Add files using upload-large-folder tool

peer_reviews/105ad8323cbc30432dfea0d327a986d4193c46b0744395233608b6bf22b098fc/supplementary_1_Peer Review Information./images_list.json

@@ -0,0 +1 @@
+[]

peer_reviews/105ad8323cbc30432dfea0d327a986d4193c46b0744395233608b6bf22b098fc/supplementary_1_Peer Review Information./supplementary_1_Peer Review Information..mmd

@@ -0,0 +1,572 @@
+
+# natureresearch  
+
+# Peer Review Information  
+
+Journal: Nature Ecology & Evolution  Manuscript Title: Palaeoecological data indicates land- use changes across Europe linked to spatial heterogeneity in mortality during the Black Death pandemic  Corresponding author name(s): Alessia Masi  
+
+## Reviewer Comments & Decisions:  
+
+# Decision Letter, initial version:  
+
+2nd August 2021  
+
+\*Please ensure you delete the link to your author homepage in this e- mail if you wish to forward it to your co- authors.  
+
+Dear Dr Masi,  
+
+Your manuscript entitled "Big Data Palaeoecology reveals significant variation in Black Death mortality in Europe" has now been seen by four reviewers, whose comments are attached. The reviewers have raised a number of concerns which will need to be addressed before we can offer publication in Nature Ecology & Evolution. We will therefore need to see your responses to the criticisms raised and to some editorial concerns, along with a revised manuscript, before we can reach a final decision regarding publication.  
+
+In particular, you will see that the reviewers emphasise the need to tone down the degree to which the narrative presented is a surprise from a historical perspective- - I don't think this compromises novelty since as reviewer 1 argues, it is valuable to show the palaeoecological dimension on land use change during the period. And as reviewer 4 comments, there is also a need to shore up the linkage between mortality and land use changes- - bringing more historical perspective in here could help.  
+
+We therefore invite you to revise your manuscript taking into account all reviewer and editor comments. Please highlight all changes in the manuscript text file [OPTIONAL: in Microsoft Word format].  
+
+We are committed to providing a fair and constructive peer- review process. Do not hesitate to contact us if there are specific requests from the reviewers that you believe are technically impossible or unlikely to yield a meaningful outcome.
+
+<--- Page Split --->
+
+# natureresearch  
+
+When revising your manuscript:  
+
+\* Include a "Response to reviewers" document detailing, point- by- point, how you addressed each reviewer comment. If no action was taken to address a point, you must provide a compelling argument. This response will be sent back to the reviewers along with the revised manuscript.  
+
+\* If you have not done so already please begin to revise your manuscript so that it conforms to our Article format instructions at http://www.nature.com/natecolevol/info/final- submission. Refer also to any guidelines provided in this letter.  
+
+\* Include a revised version of any required reporting checklist. It will be available to referees (and, potentially, statisticians) to aid in their evaluation if the manuscript goes back for peer review. A revised checklist is essential for re- review of the paper.  
+
+Please use the link below to submit your revised manuscript and related files:  
+
+## [REDACTED]  
+
+<strong>Note: </strong> This URL links to your confidential home page and associated information about manuscripts you may have submitted, or that you are reviewing for us. If you wish to forward this email to co- authors, please delete the link to your homepage.  
+
+We hope to receive your revised manuscript within four to eight weeks. If you cannot send it within this time, please let us know. We will be happy to consider your revision so long as nothing similar has been accepted for publication at Nature Ecology & Evolution or published elsewhere.  
+
+Nature Ecology & Evolution is committed to improving transparency in authorship. As part of our efforts in this direction, we are now requesting that all authors identified as 'corresponding author' on published papers create and link their Open Researcher and Contributor Identifier (ORCID) with their account on the Manuscript Tracking System (MTS), prior to acceptance. ORCID helps the scientific community achieve unambiguous attribution of all scholarly contributions. You can create and link your ORCID from the home page of the MTS by clicking on 'Modify my Springer Nature account'. For more information please visit please visit <a href="http://www.springernature.com/orcid">www.springernature.com/orcid</a>.  
+
+Please do not hesitate to contact me if you have any questions or would like to discuss these revisions further.  
+
+We look forward to seeing the revised manuscript and thank you for the opportunity to review your work.  
+
+## [REDACTED]  
+
+Reviewer expertise:  
+
+Reviewer #1: medieval economic history
+
+<--- Page Split --->
+
+# natureresearch  
+
+Reviewer #2: Black Death archaeology  
+
+Reviewer #3: palaeoecology and palynology  
+
+Reviewer #4: palaeoecology and palynology  
+
+Reviewers' comments:  
+
+Reviewer #1 (Remarks to the Author):  
+
+This is an interesting article about the regional variation of agrarian changes in Europe during the late middle ages. Probably for the first time such a large amount of data based on pollen analysis has been brought together to study the regional variation and changes of rural cultivation activity during that period.  
+
+It confirms the existing knowledge (already since Biraben's publication of 1975 and many others after that) that the influence of the late medieval crisis hit Europe with a huge regional variation. However one of the most important merits of the article is that it is the first time that a study like this is done on the basis a large scale dataset out of the positive sciences.  
+
+However, contrary to what the title and some parts of the article are suggesting, the data do not exclusively reveal the impact of the pest epidemic of 1347- 1352. They are on the contrary revealing also other changes in the late medieval society that are due to other factors as well. Indeed other demographic catastrophes were typical for the (later) middle ages as well and certainly also are reflected in pollen analyses. We think about famines (e.g. a° 1315- 1317), wars (that became much more devastating in the late middle ages and also caused epidemic diseases) but especially other epidemics and most importantly the s.c. 'echo- (pest- )epidemics' after the first plague epidemic (e.g. 1357- 9, 1373, 1400- 1401...). However also these epidemics occurred with huge regional variations. Moreover during that period structural changes of the rural society took place that were only partly linked to the black death (e.g. land concentration and the occurrence of new elites).  
+
+To conclude: Interesting article but before publication we advise that the content should be changed as mentioned and that a new title should be formulated that is more conform to its content.  
+
+Reviewer #2 (Remarks to the Author):  
+
+This interesting study uses changes in the type of pollen in sediment cores to investigate the consequences of the black death upon the types of plant growing. The study uses a logical approach, and the illustrations explain the data convincingly.  
+
+Estimating Black Death Mortality section (lines 131- 165). You write that until now texts have been the best way to study the impact of the Black Death on population size, although it is good that you do mention papers about changing pollution levels. However, other techniques are available. One study looked at the number of fragments of pottery in test pits to show that there was a significant drop in
+
+<--- Page Split --->
+
+# natureresearch  
+
+pottery use and production in England following the Black Death, likely a result of the smaller population size. I would encourage you to read this and give a couple of sentences on its approach (it is not my paper): Lewis, C. (2016) Disaster recovery: new archaeological evidence for the long- term impact of the 'calamitous' fourteenth century. Antiquity 90(351): 777- 797.  
+
+Minor point: Abstract first sentence: 'reknown' should read 'renown'.  
+
+Reviewer #3 (Remarks to the Author):  
+
+This is a very interesting and topical paper. In principle, I was surprised by the methodological approach, but after studying it repeatedly, I find that the idea is provocative and has great explanatory power. The conclusion about the spatial and social heterogeneity of the epidemic is common sense regardless of conventional historicist approaches based on inevitably biased sources.  
+
+Obviously, pollen analysis also entails its spatial sampling, analysis and interpretation biases, in addition to the biases derived from the resolution in the identification of palynological types. Not even close to all the pollen analyses published in the studied territory are presented here either. Naturally, all these limitations are unapproachable in the current state of collaboration with Big Data.  
+
+That said, the epistemological and methodological approach of this study seems fresh and healthy to me as a pioneering study on the frontiers of knowledge. A much needed consilience between the natural sciences and the social sciences. A brilliant work in my opinion, which will be a must- read for the general public and which I predict a considerable media success.  
+
+My enthusiasm cannot but lead to the belief that the paper can be published in its current form. The language is accessible, the information well collected and the message clear and direct, without frills. Congratulations to the authors.  
+
+Reviewer #4 (Remarks to the Author):  
+
+In "Big Data Palaeoecology reveals significant variation in Black Death mortality in Europe" the authors evaluate palaeorecord of series of sedimentary sequences from 19 European countries to evaluate whether landscape changes could be linked with the hypothesis that half of the population died within a single year. This is a very interesting paper with an interesting approach and I have enjoyed reading it. It is really exciting seeing how fossil pollen time- series are answering big and relevant questions of concern to society as pandemics are. However, I have some concerns I would like the authors to consider.  
+
+Main points:  
+
+1) Mortality and landscape change: After reading the paper with interest and SI (lines 445 474) the
+
+<--- Page Split --->
+
+# natureresearch  
+
+"direct" link between population decline or mortality and landscape change that I think is the main hypothesis of the paper is not clear to me. The authors used \(>1600\) palynological time- series to evaluate the demographic impact of the Black Death on a regional scale across Europe. I think that the results of the paper are not strictly quantifying mortality (e.g. what magnitude?) but responses of local/regional vegetation/landscapes to changes on land- use, and other factors including mortality. However, I think the results are still relevant.  
+
+I also found that the manuscript requires an improved discussion on demographic changes- how the authors interpret demographic changes in their results? It is mortality quantified "only" with th increase and decrease of cereals, etc? In Lines 191- 198 there are more discussion on validation. And I thank the authors for including this but I keep struggling in understanding what is the actual quantification of mortality. Also, why Poland experienced growth? More is required in here.  
+
+2) Time period: It requires more detail on why the analysis are focusing on the period between 1250 and 1450 CE. I think this detail should be explored in-depth to better discuss the otherwise relevant results.  
+
+3) Regional selection: Another issues I think it requires more description/discussion is why the authors included and excluded sites (Fig 1). I understand the limitation on finding good quality datasets for the time period of interest but I think that comparing sites located in mountain areas vs coastal areas and/or with different population densities, etc. probably impacted the findings.  
+
+Minor points  
+
+Line 113: I think it is missing the years that the pandemic was active. Please add the years.  
+
+Ecological indices: It is missing a line with the relevance of these indices to help linking landscape change and the scale of the Black Death mortality. Why are these indices important to answer the question? I read the detail in SI but still not clear to me why the authors are using them. Not suggesting that the approach is not relevant but there are so many analysis to quantify how cereals and different vegetation types change over time that it require a better justification.  
+
+Lines 171- 172: The authors suggested that "Pollen data can be used to assess past demographic dynamics as human pressure on the landscape in the preindustrial period was directly dependent on the availability of rural labour." I think this line is a bit confusing as it is mixing up to rural vs city demographics with the main point of the paper that is mortality.  
+
+Line 261: This is a very strong sentence "The significant variability of Black Death mortality that our BDP approach identifies remains..." I'm struggling to accept that the results are showing differences in mortality rates. I think that it is an interesting hypothesis but it could well be that the landscape change found may respond to differences in vegetation resilience, different baseline population densities and cultures, land- use, culture differences, etc.  
+
+Lines 286- 295 probably could be placed at the beginning of the manuscript to help integrating the spatial distribution of the Black Death and to better justify the time- period.  
+
+Lines 487- 494: Depth age models- have the authors re- calibrated them using the new calibration curves?
+
+<--- Page Split --->
+
+# natureresearch  
+
+I hope my comments are helpful to improve the manuscript.  
+
+\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*/\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*END\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*/\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\*  
+
+# Author Rebuttal to Initial comments  
+
+## Response to reviewers  
+
+## Big Data Palaeoecology reveals significant variation in Black Death mortality in Europe  
+
+Izdebski et al.  
+
+Editor: In particular, you will see that the reviewers emphasise the need to tone down the degree to which the narrative presented is a surprise from a historical perspective- - I don't think this compromises novelty since as reviewer 1 argues, it is valuable to show the palaeoecological dimension on land use change during the period.  
+
+- > We toned down the first section of the article and took into account the non-Anglophone historiography, which contrary to the recent English-speaking consensus on the Black Death as a universal catastrophic event, noticed some regional variation during the initial phases of the second plague pandemic.  
+
+And as reviewer 4 comments, there is also a need to shore up the linkage between mortality and land use changes- - bringing more historical perspective in here could help.  
+
+- > We rewrote the relevant (first) section of the Methods, providing it with additional references. In particular, we divided this section into two clearly distinguished main points: (1) the close connection between landscape changes as revealed by palynology and population dynamics as revealed by historical sources, as proved by several recent historical-palynological studies; (2)
+
+<--- Page Split --->
+
+# natureresearch  
+
+the landscape changes to be expected in plague- struck areas during the Black Death, based on the current consensus on estimated population decline among the historians of the Black Death.  
+
+Reviewer expertise:  
+
+Reviewer #1: medieval economic history  
+
+Reviewer #2: Black Death archaeology  
+
+Reviewer #3: palaeoecology and palynology  
+
+Reviewer #4: palaeoecology and palynology  
+
+Reviewer #1 (Remarks to the Author):  
+
+This is an interesting article about the regional variation of agrarian changes in Europe during the late middle ages. Probably for the first time such a large amount of data based on pollen analysis has been brought together to study the regional variation and changes of rural cultivation activity during that period.  
+
+It confirms the existing knowledge (already since Biraben's publication of 1975 and many others after that) that the influence of the late medieval crisis hit Europe with a huge regional variation. However one of the most important merits of the article is that it is the first time that a study like this is done on the basis a large scale dataset out of the positive sciences.  
+
+However, contrary to what the title and some parts of the article are suggesting, the data do not exclusively reveal the impact of the pest epidemic of 1347- 1352. They are on the contrary revealing also other changes in the late medieval society that are due to other factors as well. Indeed other demographic catastrophes were typical for the (later) middle ages as well and certainly also are reflected in pollen analyses. We think about famines (e.g. a° 1315- 1317), wars (that became much more devastating in the late middle ages and also caused epidemic
+
+<--- Page Split --->
+
+# natureresearch  
+
+diseases) but especially other epidemics and most importantly the s.c. 'echo- (pest- )epidemics' after the first plague epidemic (e.g. 1357- 9, 1373, 1400- 1401...). However also these epidemics occurred with huge regional variations. Moreover during that period structural changes of the rural society took place that were only partly linked to the black death (e.g. land concentration and the occurrence of new elites).  
+
+- > We are grateful for these careful and constructive comments. We had long discussions among the historians who co-authored this paper about the extent to which our results are connected with phenomena other than the Black Death itself. Our conclusion is that our method produces results that actually are connected directly to the Black Death understood as the initial introduction of Y. pestis to Europe during the second pandemic. Of course, the later echo-epidemics (later outbreaks) amplified the consequences of the Black Death, but the role of the outbreak of the Black Death was crucial for whether a population collapse - leading to significant landscape changes - occurred in a particular region or not. It is the Black Death itself, the pandemic of the mid-fourteenth-century, that is attributed with killing half of the population of Europe in a few years, not the subsequent outbreaks, which only added, albeit in a regionalized way, to the mortality of the Black Death decades after the Black Death. As such, it is the Black Death that we focus on. It was the major demographic moment of change in fourteenth-century European history and, in fact, in late medieval history more generally. Very few historians would expect the landscape of a region to have been altered significantly if it was not struck by the Black Death, but only by secondary and tertiary outbreaks of plague after the Black Death in the fourteenth century.  
+
+To these remark, we should add that our method is not focused on showing the different trajectories of the late medieval crisis (as was, for instance, Izdebski et al. 2016, https://doi.org/10.1016/j.eeh.2015.10.003); instead of focusing on time series and trend analysis, we focus on the extent to which a single event, well- constrained in time - the Black Death - can be associated with landscape changes across Europe. It may well be that all the other phenomena that are collectively known as the late medieval crisis left their imprint on the landscape as well (for instance, our data show the impact of the Great Famine of 1315- 1317 in Devon/England, for which we have no significant impact of the Black Death), but our method in this particular article is set up in such a way that it does not reveal them: it is very clearly focused on the Black Death.  
+
+Hence, the results presented in our article concern 100- year periods before and after the Black Death, that is 4 human generations (the minimum time to overcome a sudden population loss in the range of 30- 50%). Additionally, in the Supplementary Material, we already showed the results for 50 and 25 year periods before and after the Black Death: they
+
+<--- Page Split --->
+
+# natureresearch  
+
+are very similar to the results of the 100 year periods, as we discussed in the main text. To emphasise this fact, and further demonstrate the focus of our analysis on the Black Death in particular, we now added additional figures to the text and commented on them. Based on the consistency between these different periods of analysis, we are convinced that the landscape changes we analyse across Europe are related directly to the Black Death, and not - as the reviewer suggests - to the 'Dantean Anomaly' and the Great Famine of the 1310s or to the late fourteenth or early fifteenth century- plague outbreaks. The impact of these phenomena is not captured by the 25 year or even 50 year periods of analysis). Thus, given the current consensus hypothesis we want to test (universal mass mortality), we are able to demonstrate that whereas it did occur in some areas (e.g. Greece, Sweden or Italy), it did not occur elsewhere (e.g. Poland or Spain). We believe this is the main merit of our article, and to some extent it is indeed confirmation of historical works of Biraben, Malwost or Gottfried, who emphasised regional differentiation of the late medieval crisis in Europe, a minority perspective today which we now make clear reference to in the article).  
+
+To conclude: Interesting article but before publication we advise that the content should be changed as mentioned and that a new title should be formulated that is more conform to its content.  
+
+- > We have suggested a new title, clearly focused on the vision of the Black Death that dominates plague studies and the public imagination - the vision which we falsify in our study.  
+
+Reviewer #2 (Remarks to the Author):  
+
+This interesting study uses changes in the type of pollen in sediment cores to investigate the consequences of the black death upon the types of plant growing. The study uses a logical approach, and the illustrations explain the data convincingly.  
+
+Estimating Black Death Mortality section (lines 131- 165). You write that until now texts have been the best way to study the impact of the Black Death on population size, although it is good that you do mention papers about changing pollution levels. However, other techniques are available. One study looked at the number of fragments of pottery in test pits to show that there was a significant drop in pottery use and production in England following the Black Death, likely a result of the smaller population size. I would encourage you to read this and give a couple of sentences on its approach (it is not my paper): Lewis, C. (2016) Disaster recovery: new archaeological evidence for the long- term impact of the 'calamitous' fourteenth century. Antiquity 90(351): 777- 797.
+
+<--- Page Split --->
+
+# natureresearch  
+
+- > We have now included this study in our method section. We also would like to thank the reviewer 2 for the positive comments on our paper and for expanding even further our interdisciplinarity.  
+
+Minor point: Abstract first sentence: 'reknown' should read 'renown'.  
+
+- > Corrected, thank you.  
+
+Reviewer #3 (Remarks to the Author):  
+
+This is a very interesting and topical paper. In principle, I was surprised by the methodological approach, but after studying it repeatedly, I find that the idea is provocative and has great explanatory power. The conclusion about the spatial and social heterogeneity of the epidemic is common sense regardless of conventional historicist approaches based on inevitably biased sources.  
+
+Obviously, pollen analysis also entails its spatial sampling, analysis and interpretation biases, in addition to the biases derived from the resolution in the identification of palynological types. Not even close to all the pollen analyses published in the studied territory are presented here either. Naturally, all these limitations are unapproachable in the current state of collaboration with Big Data.  
+
+That said, the epistemological and methodological approach of this study seems fresh and healthy to me as a pioneering study on the frontiers of knowledge. A much needed consilience between the natural sciences and the social sciences. A brilliant work in my opinion, which will be a must- read for the general public and which I predict a considerable media success.  
+
+My enthusiasm cannot but lead to the belief that the paper can be published in its current form. The language is accessible, the information well collected and the message clear and direct, without frills. Congratulations to the authors.  
+
+- > Thank you very much for such an encouraging review!  
+
+Reviewer #4 (Remarks to the Author):
+
+<--- Page Split --->
+
+# natureresearch  
+
+In "Big Data Palaeoecology reveals significant variation in Black Death mortality in Europe" the authors evaluate paleorecord of series of sedimentary sequences from 19 European countries to evaluate whether landscape changes could be linked with the hypothesis that half of the population died within a single year. This is a very interesting paper with an interesting approach and I have enjoyed reading it. It is really exciting seeing how fossil pollen time- series are answering big and relevant questions of concern to society as pandemics are. However, I have some concerns I would like the authors to consider.  
+
+Main points:  
+
+1) Mortality and landscape change: After reading the paper with interest and SI (lines 445 474) the "direct" link between population decline or mortality and landscape change that I think is the main hypothesis of the paper is not clear to me. The authors used >1600 palynological time-series to evaluate the demographic impact of the Black Death on a regional scale across Europe. I think that the results of the paper are not strictly quantifying mortality (e.g. what magnitude?) but responses of local/regional vegetation/landscapes to changes on land-use, and other factors including mortality. However, I think the results are still relevant.  
+
+-> Yes, indeed. In agreement with the comments of R1, and by changing the title, we now tried to make clear that while we do study landscape changes (or lack thereof), we do so entirely in order to test the currently dominant view that the Black Death led to universal mortality within the range of 30-50%. We now make clear in the methods section why this is important to us: such a truly significant mortality would lead to profound and sudden landscape change, which we indeed observe in some of our study regions.  
+
+I also found that the manuscript requires an improved discussion on demographic changes- how the authors interpret demographic changes in their results? It is mortality quantified "only" with th increase and decrease of cereals, etc? In Lines 191- 198 there are more discussion on validation. And I thank the authors for including this but I keep struggling in understanding what is the actual quantification of mortality.  
+
+-> Entirely understood. We have now improved our discussion of the connection to plague mortality in the relevant sections of the Methods section and in the main text. As we explain, we do not provide direct, concrete estimates of mortality, as the state of the research on palynology and historical demography has not yet reached this stage, despite some major
+
+<--- Page Split --->
+
+# natureresearch  
+
+recent advances in this respect. Our aim instead is to use pollen data to verify the popular vision of the Black Death as a universal mortality catastrophe in Europe, leading to \(30 - 50\%\) of the population dying in a few mere years. As we now try to explain more clearly, such a scenario would result in major landscape changes following the sudden and profound reduction of human pressure on the landscape. We observe significant population erosion in some parts of Europe (consistently with the limited written evidence available for Black Death mortality), while not in others (for which we provide explanation in the concluding section of the main text).  
+
+Also, why Poland experienced growth? More is required in here.  
+
+- > We now explain this fact.  
+
+2) Time period: It requires more detail on why the analysis are focusing on the period between 1250 and 1450 CE. I think this detail should be explored in-depth to better discuss the otherwise relevant results.  
+
+- > We now discuss our choice of 100 year slices of time on either side of 1350 as our main period of analysis in the first section of the Methods. We now also discuss the results we obtained for shorter time periods in the main text.  
+
+3) Regional selection: Another issues I think it requires more description/discussion is why the authors included and excluded sites (Fig 1). I understand the limitation on finding good quality datasets for the time period of interest but I think that comparing sites located in mountain areas vs coastal areas and/or with different population densities, etc. probably impacted the findings.  
+
+- > We now provide two additional supplementary figures that show the geographical setting (lowland vs highland) of our sites as well as the population densities around our sites in the fourteenth century. These variables are not linked to the results we obtained for our BDP pollen indicators, however. Please note that we undertook as well a comparative study of change in the landscape over periods of 200, 100 and 50 years. We assume that over such short time scales physical geographical characteristics of each site remain stable and do not impact our results and conclusions.
+
+<--- Page Split --->
+
+# natureresearch  
+
+Minor points  
+
+Line 113: I think it is missing the years that the pandemic was active. Please add the years.  
+
+- > This Black Death pandemic probably began in the 1330s in Central Asia and in many parts of Eurasia and Africa - outside of Europe - continued until the late 1350s. For this reason, we believe providing an approximate chronological range of the "mid-14th c." is most correct. For the entire Eurasia and Africa, providing set years would be difficult to substantiate and there is no consensus (contrary to the dates for Europe only).  
+
+Ecological indices: It is missing a line with the relevance of these indices to help linking landscape change and the scale of the Black Death mortality. Why are these indices important to answer the question? I read the detail in SI but still not clear to me why the authors are using them. Not suggesting that the approach is not relevant but there are so many analysis to quantify how cereals and different vegetation types change over time that it require a better justification.  
+
+- > We have now provided more explanation in the main text and in the Methods.  
+
+Lines 171- 172: The authors suggested that "Pollen data can be used to assess past demographic dynamics as human pressure on the landscape in the preindustrial period was directly dependent on the availability of rural labour." I think this line is a bit confusing as it is mixing up to rural vs city demographics with the main point of the paper that is mortality.  
+
+- > As we explain in the Methods, \(\sim 90\%\) of the late medieval European population lived in the countryside, so in order to understand overall mortality levels during the Black Death, our primary focus should be the rural populations.  
+
+Line 261: This is a very strong sentence "The significant variability of Black Death mortality that our BDP approach identifies remains..." I'm struggling to accept that the results are showing differences in mortality rates. I think that it is an interesting hypothesis but it could well be that
+
+<--- Page Split --->
+
+# natureresearch  
+
+the landscape change found may respond to differences in vegetation resilience, different baseline population densities and cultures, land- use, culture differences, etc.  
+
+- > Duly noted. As we argue in the Methods, though, the changes we observe are related to human pressure on the landscape, which in turn was dependent on population levels. We also focus primarily on cultivated plants, cereals, whose presence in the landscape diminishes dramatically and over very short time scales without human activity/pressure that supports these plants, as we explain in the Methods. Moreover, thanks to our strict chronological focus, also as now explained further in response to the comments of R1, we are in the position to argue that these landscape changes (or lack thereof) were directly related to the Black Death, that is, the commonly supposed most important demographic event of the fourteenth century. At the same time, we agree with R4 that the factors s/he mentions were important as moderating the Black Death's impact at the regional and local level - and we discuss this point in the concluding section of the main text.  
+
+Lines 286- 295 probably could be placed at the beginning of the manuscript to help integrating the spatial distribution of the Black Death and to better justify the time- period.  
+
+- > We now focus more on explaining the temporal and spatial focus we have adopted throughout the text.  
+
+Lines 487- 494: Depth age models- have the authors re- calibrated them using the new calibration curves?  
+
+- > Yes, absolutely: as we explain in the Methods, we re- calculated age-depth models for a large portion of our sites: especially for those that were using the older calibration curves.  
+
+I hope my comments are helpful to improve the manuscript.  
+
+- > Yes, indeed, thank you very much!
+
+<--- Page Split --->
+
+# natureresearch  
+
+## Decision Letter, first revision:  
+
+8th October 2021  
+
+Dear Dr. Masi,  
+
+Thank you for submitting your revised manuscript "The Black Death Killed Far Fewer Europeans than Commonly Thought, Big Data Paleoecology Reveals" (NATECoLEVEL- 210614034A). It has now been seen again by the original reviewers and their comments are below. The reviewers find that the paper has improved in revision, and therefore we'll be happy in principle to publish it in Nature Ecology & Evolution, pending revisions to satisfy the reviewers' final requests and to comply with our editorial and formatting guidelines. As explained below, our full recommendations will be sent in a later email, but for now, please note that both reviewers recommend that considerable nuance be brought into the discussion--your main text is on the short side by my rough calculation, so I estimate that you should be able to incorporate up to about 500 words of additional discussion which should address the remaining reviewer concerns.  
+
+If the current version of your manuscript is in a PDF format, please email us a copy of the file in an editable format (Microsoft Word or LaTex)-- we can not proceed with PDFs at this stage.  
+
+We are now performing detailed checks on your paper and will send you a checklist detailing our editorial and formatting requirements in about a week. Please do not upload the final materials and make any revisions until you receive this additional information from us.  
+
+Thank you again for your interest in Nature Ecology & Evolution. Please do not hesitate to contact me if you have any questions.  
+
+## [REDACTED]  
+
+Reviewer #1 (Remarks to the Author):  
+
+## GENERAL COMMENT:  
+
+The article has been upgraded a lot: the amount of relevant literature has been increased and the text is more nuanced.  
+
+A few remarks anyway (put in the reply- text of the authors)  
+
+## DETAILS:  
+
+Reply text of the authors to the first remarks:  
+
+"Very few historians would expect the landscape of a region to have been altered significantly if it was not struck by the Black Death, but only by secondary and tertiary outbreaks of plague after the Black Death in the fourteenth century."  
+
+Reply of the reviewer: Anyway there were important regions where the later outbreaks had much more influence that the first outbreak. Moreover if there would not have been 'echo- epidemics', the economy (and the landscapes) probably would
+
+<--- Page Split --->
+
+# natureresearch  
+
+have recovered early.  
+
+Reply text of the authors to the review  
+
+"To these remark, we should add that our method is not focused on showing the different trajectories of the late medieval crisis (as was, for instance, Izdebski et al. 2016, https://doi.org/10.1016/j.eeh.2015.10.003); instead of focusing on time series and trend analysis, we focus on the extent to which a single event, well constrained in time - the Black Death - can be associated with landscape changes across Europe. It may well be that all the other phenomena that are collectively known as the late medieval crisis left their imprint on the landscape as well (for instance, our data show the impact of the Great Famine of 1315- 1317 in Devon/England, for which we have no significant impact of the Black Death), but our method in this particular article is set up in such a way that it does not reveal them: it is very clearly focused on the Black Death."  
+
+Reply of the reviewer:  
+
+My remark remains...Why should the data reveal more the influence of the black dead than other elements of the rural economy that were changing in the later middle ages? Of course it is logic that one cannot deal with these other elements but at least it could be mentioned.  
+
+Reviewer #4 (Remarks to the Author):  
+
+This version of the manuscript reads very well and I thank the authors for taking into consideration my comments. This paper is interesting, well written, with fantastic figures, and can bring good discussion on how to integrate data from different paleo- sciences. I agree that population changes may result in major landscape changes (accounted using pollen) following the sudden and profound reduction of human pressure on the landscape that the BD produced- we have a good example with COVID- 19. However, I keep struggling with the concept that the land- use changes found are reflecting differences in mortality rates or the magnitude of the severity of BD, only (line 313, title, etc).  
+
+Overall, the authors should tone down a bit more the narrative around novelty (e.g. pioneering). Using pollen data (or Big- data palaeoecology as the authors refers too) to infer land- use changes or landscape changes is not novel as it is not the link between land- use change and population dynamics. What is novel and exciting of this paper is the integration of historical and paleoecological sources to answer a current important research question: what happen to the landscape when there is a global pandemic with all the associated factors (economical crisis, rates of mortality, duration of the pandemic, type of vegetation, etc...)?  
+
+Interesting paper!
+
+<--- Page Split --->
+
+# natureresearch  
+
+13th October 2021  
+
+Dear Dr. Masi,  
+
+Thank you for your patience as we've prepared the guidelines for final submission of your Nature Ecology & Evolution manuscript, "The Black Death Killed Far Fewer Europeans than Commonly Thought, Big Data Paleoecology Reveals" (NATECOLEVOL- 210614034A). Please carefully follow the step- by- step instructions provided in the attached file, and add a response in each row of the table to indicate the changes that you have made. Please also check and comment on any additional marked- up edits we have proposed within the text. Ensuring that each point is addressed will help to ensure that your revised manuscript can be swiftly handed over to our production team.  
+
+\*\*We would like to start working on your revised paper, with all of the requested files and forms, as soon as possible (preferably within two weeks). Please get in contact with us immediately if you anticipate it taking more than two weeks to submit these revised files.\*\*  
+
+When you upload your final materials, please include a point- by- point response to any remaining reviewer comments.  
+
+If you have not done so already, please alert us to any related manuscripts from your group that are under consideration or in press at other journals, or are being written up for submission to other journals (see: https://www.nature.com/nature- research/editorial- policies/plagiarism#policy- on- duplicate- publication for details).  
+
+In recognition of the time and expertise our reviewers provide to Nature Ecology & Evolution's editorial process, we would like to formally acknowledge their contribution to the external peer review of your manuscript entitled "The Black Death Killed Far Fewer Europeans than Commonly Thought, Big Data Paleoecology Reveals". For those reviewers who give their assent, we will be publishing their names alongside the published article.  
+
+Nature Ecology & Evolution offers a Transparent Peer Review option for new original research manuscripts submitted after December 1st, 2019. As part of this initiative, we encourage our authors to support increased transparency into the peer review process by agreeing to have the reviewer comments, author rebuttal letters, and editorial decision letters published as a Supplementary item. When you submit your final files please clearly state in your cover letter whether or not you would like to participate in this initiative. Please note that failure to state your preference will result in delays in accepting your manuscript for publication.  
+
+<b>Cover suggestions</b>  
+
+As you prepare your final files we encourage you to consider whether you have any images or illustrations that may be appropriate for use on the cover of Nature Ecology & Evolution.  
+
+Covers should be both aesthetically appealing and scientifically relevant, and should be supplied at the best quality available. Due to the prominence of these images, we do not generally select images featuring faces, children, text, graphs, schematic drawings, or collages on our covers.
+
+<--- Page Split --->
+
+# natureresearch  
+
+We accept TIFF, JPEG, PNG or PSD file formats (a layered PSD file would be ideal), and the image should be at least 300ppi resolution (preferably 600- 1200 ppi), in CMYK colour mode.  
+
+If your image is selected, we may also use it on the journal website as a banner image, and may need to make artistic alterations to fit our journal style.  
+
+Please submit your suggestions, clearly labeled, along with your final files. We'll be in touch if more information is needed.  
+
+Nature Ecology & Evolution has now transitioned to a unified Rights Collection system which will allow our Author Services team to quickly and easily collect the rights and permissions required to publish your work. Approximately 10 days after your paper is formally accepted, you will receive an email in providing you with a link to complete the grant of rights. If your paper is eligible for Open Access, our Author Services team will also be in touch regarding any additional information that may be required to arrange payment for your article.  
+
+Please note that <i>Nature Ecology & Evolution</i> is a Transformative Journal (TJ). Authors may publish their research with us through the traditional subscription access route or make their paper immediately open access through payment of an article- processing charge (APC). Authors will not be required to make a final decision about access to their article until it has been accepted. <a href="https://www.springernature.com/gp/open- research/transformative- journals"> Find out more about Transformative Journals</a>  
+
+<B>Authors may need to take specific actions to achieve <a href="https://www.springernature.com/gp/open- research/funding/policy- compliance- faqs"> compliance</a> with funder and institutional open access mandates.</b> For submissions from January 2021, if your research is supported by a funder that requires immediate open access (e.g. according to <a href="https://www.springernature.com/gp/open- research/plan- s- compliance">Plan S principles</a>) then you should select the gold OA route, and we will direct you to the compliant route where possible. For authors selecting the subscription publication route our standard licensing terms will need to be accepted, including our <a href="https://www.springernature.com/gp/open- research/policies/journal- policies">self- archiving policies</a>. Those standard licensing terms will supersede any other terms that the author or any third party may assert apply to any version of the manuscript.  
+
+Please note that you will not receive your proofs until the publishing agreement has been received through our system.  
+
+For information regarding our different publishing models please see our <a href="https://www.springernature.com/gp/open- research/transformative- journals"> Transformative Journals </a> page. If you have any questions about costs, Open Access requirements, or our legal forms, please contact ASJournals@springernature.com.  
+
+Please use the following link for uploading these materials:
+
+<--- Page Split --->
+
+# natureresearch  
+
+## [REDACTED]  
+
+If you have any further questions, please feel free to contact me.  
+
+## [REDACTED]  
+
+Reviewer #1: Remarks to the Author: GENERAL COMMENT:  
+
+The article has been upgraded a lot: the amount of relevant literature has been increased and the text is more nuanced.  
+
+A few remarks anyway (put in the reply- text of the authors)  
+
+DETAILS:  
+
+Reply text of the authors to the first remarks:  
+
+"Very few historians would expect the landscape of a region to have been altered significantly if it was not struck by the Black Death, but only by secondary and tertiary outbreaks of plague after the Black Death in the fourteenth century."  
+
+Reply of the reviewer: Anyway there were important regions where the later outbreaks had much more influence that the first outbreak. Moreover if there would not have been 'echoepidemics', the economy (and the landscapes) probably would have recovered early.  
+
+Reply text of the authors to the review  
+
+"To these remark, we should add that our method is not focused on showing the different trajectories of the late medieval crisis (as was, for instance, Izdebski et al. 2016, https://doi.org/10.1016/j.eeh.2015.10.003); instead of focusing on time series and trend analysis, we focus on the extent to which a single event, well constrained in time - the Black Death - can be associated with landscape changes across Europe. It may well be that all the other phenomena that are collectively known as the late medieval crisis left their imprint on the landscape as well (for instance, our data show the impact of the Great Famine of 1315- 1317 in Devon/England, for which we have no significant impact of the Black Death), but our method in this particular article is set up in such a way that it does not reveal them: it is very clearly focused on the Black Death."  
+
+Reply of the reviewer:  
+
+My remark remains...Why should the data reveal more the influence of the black dead than other elements of the rural economy that were changing in the later middle ages? Of course it is logic that one cannot deal with these other elements but at least it could be mentioned.  
+
+Reviewer #4: Remarks to the Author:
+
+<--- Page Split --->
+
+# natureresearch  
+
+This version of the manuscript reads very well and I thank the authors for taking into consideration my comments. This paper is interesting, well written, with fantastic figures, and can bring good discussion on how to integrate data from different paleo- sciences. I agree that population changes may result in major landscape changes (accounted using pollen) following the sudden and profound reduction of human pressure on the landscape that the BD produced- we have a good example with COVID- 19. However, I keep struggling with the concept that the land- use changes found are reflecting differences in mortality rates or the magnitude of the severity of BD, only (line 313, title, etc).  
+
+Overall, the authors should tone down a bit more the narrative around novelty (e.g. pioneering). Using pollen data (or Big- data palaeoecology as the authors refers too) to infer land- use changes or landscape changes is not novel as it is not the link between land- use change and population dynamics. What is novel and exciting of this paper is the integration of historical and paleoecological sources to answer a current important research question: what happen to the landscape when there is a global pandemic with all the associated factors (economical crisis, rates of mortality, duration of the pandemic, type of vegetation, etc...)?  
+
+Interesting paper!  
+
+Author Rebuttal, first revision:  
+
+## Response to the reviewers' comments (second revision)  
+
+Reviewer #1: Remarks to the Author: GENERAL COMMENT: The article has been upgraded a lot: the amount of relevant literature has been increased and the text is more nuanced.  
+
+- > Thank you!  
+
+A few remarks anyway (put in the reply- text of the authors)  
+
+DETAILS:  
+
+Reply text of the authors to the first remarks:  
+
+"Very few historians would expect the landscape of a region to have been altered significantly if it was not struck by the Black Death, but only by secondary and tertiary outbreaks of plague after the Black Death in the fourteenth century."  
+
+Reply of the reviewer: Anyway there were important regions where the later outbreaks had much more influence that the first outbreak. Moreover if there would not have been 'echo
+
+<--- Page Split --->
+
+# natureresearch  
+
+epidemics', the economy (and the landscapes) probably would have recovered early.  
+
+- > We took account of this point in the discussion.  
+
+Reply text of the authors to the review  
+
+"To these remark, we should add that our method is not focused on showing the different trajectories of the late medieval crisis (as was, for instance, Izdebski et al. 2016, https://doi.org/10.1016/j.eeh.2015.10.003); instead of focusing on time series and trend analysis, we focus on the extent to which a single event, well constrained in time - the Black Death - can be associated with landscape changes across Europe. It may well be that all the other phenomena that are collectively known as the late medieval crisis left their imprint on the landscape as well (for instance, our data show the impact of the Great Famine of 1315- 1317 in Devon/England, for which we have no significant impact of the Black Death), but our method in this particular article is set up in such a way that it does not reveal them: it is very clearly focused on the Black Death."  
+
+Reply of the reviewer:  
+
+My remark remains...Why should the data reveal more the influence of the black dead than other elements of the rural economy that were changing in the later middle ages? Of course it is logic that one cannot deal with these other elements but at least it could be mentioned.  
+
+- > As suggested, we now discuss these elements.  
+
+Reviewer #4:  
+
+Remarks to the Author:  
+
+This version of the manuscript reads very well and I thank the authors for taking into consideration my comments. This paper is interesting, well written, with fantastic figures, and can bring good discussion on how to integrate data from different paleo- sciences. I agree that population changes may result in major landscape changes (accounted using pollen) following the sudden and profound reduction of human pressure on the landscape that the BD produced- we have a good example with COVID- 19. However, I keep struggling with the concept that the land- use changes found are reflecting differences in mortality rates or the magnitude of the severity of BD, only (line 313, title, etc).  
+
+- > We modified the text in order to account for the context-dependent complexities involved in this potential link.  
+
+Overall, the authors should tone down a bit more the narrative around novelty (e.g. pioneering). Using pollen data (or Big- data palaeoecology as the authors refers too) to infer land- use changes or landscape changes is not novel as it is not the link between land- use change and population dynamics. What is novel and exciting of this paper is the integration of historical and paleoecological sources to answer a current important research question: what
+
+<--- Page Split --->
+
+# natureresearch  
+
+happen to the landscape when there is a global pandemic with all the associated factors (economical crisis, rates of mortality, duration of the pandemic, type of vegetation, etc...)?  
+
+- > Thank you. We now try to make it clear – and the narrative has been toned down.  
+
+Interesting paper!  
+
+- > Thank you!  
+
+## Final Decision Letter:  
+
+Dear Alessia,  
+
+We are pleased to inform you that your Article entitled "Palaeoecological data indicates land- use changes across Europe linked to spatial heterogeneity in mortality during the Black Death pandemic", has now been accepted for publication in Nature Ecology & Evolution.  
+
+Over the next few weeks, your paper will be copyedited to ensure that it conforms to Nature Ecology and Evolution style. Once your paper is typeset, you will receive an email with a link to choose the appropriate publishing options for your paper and our Author Services team will be in touch regarding any additional information that may be required  
+
+After the grant of rights is completed, you will receive a link to your electronic proof via email with a request to make any corrections within 48 hours. If, when you receive your proof, you cannot meet this deadline, please inform us at rjsproduction@springernature.com immediately.  
+
+You will not receive your proofs until the publishing agreement has been received through our system  
+
+Due to the importance of these deadlines, we ask you please us know now whether you will be difficult to contact over the next month. If this is the case, we ask you provide us with the contact information (email, phone and fax) of someone who will be able to check the proofs on your behalf, and who will be available to address any last- minute problems. Once your paper has been scheduled for online publication, the Nature press office will be in touch to confirm the details.  
+
+Acceptance of your manuscript is conditional on all authors' agreement with our publication policies (see www.nature.com/authors/policies/index.html). In particular your manuscript must not be published elsewhere and there must be no announcement of the work to any media outlet until the publication date (the day on which it is uploaded onto our web site).  
+
+Please note that <i>Nature Ecology & Evolution</i> is a Transformative Journal (TJ). Authors may publish their research with us through the traditional subscription access route or make their paper immediately open access through payment of an article- processing charge (APC). Authors will not be
+
+<--- Page Split --->
+
+# natureresearch  
+
+required to make a final decision about access to their article until it has been accepted. <a href="https://www.springernature.com/gp/open- research/transformative- journals"> Find out more about Transformative Journals</a>  
+
+<B>Authors may need to take specific actions to achieve <a href="https://www.springernature.com/gp/open- research/funding/policy- compliance- faqs"> compliance</a> with funder and institutional open access mandates.</b> For submissions from January 2021, if your research is supported by a funder that requires immediate open access (e.g. according to <a href="https://www.springernature.com/gp/open- research/plan- s- compliance">Plan S principles</a>) then you should select the gold OA route, and we will direct you to the compliant route where possible. For authors selecting the subscription publication route our standard licensing terms will need to be accepted, including our <a href="https://www.springernature.com/gp/open- research/policies/journal- policies">self- archiving policies</a>. Those standard licensing terms will supersede any other terms that the author or any third party may assert apply to any version of the manuscript.  
+
+In approximately 10 business days you will receive an email with a link to choose the appropriate publishing options for your paper and our Author Services team will be in touch regarding any additional information that may be required.  
+
+You will not receive your proofs until the publishing agreement has been received through our system.  
+
+If you have any questions about our publishing options, costs, Open Access requirements, or our legal forms, please contact ASJournals@springernature.com  
+
+An online order form for reprints of your paper is available at <a  
+
+href="https://www.nature.com/reprints/author- reprints.html">https://www.nature.com/reprints/author- reprints.html</a>. All co- authors, authors' institutions and authors' funding agencies can order reprints using the form appropriate to their geographical region.  
+
+We welcome the submission of potential cover material (including a short caption of around 40 words) related to your manuscript; suggestions should be sent to Nature Ecology & Evolution as electronic files (the image should be 300 dpi at \(210 \times 297 \text{mm}\) in either TIFF or JPEG format). Please note that such pictures should be selected more for their aesthetic appeal than for their scientific content, and that colour images work better than black and white or grayscale images. Please do not try to design a cover with the Nature Ecology & Evolution logo etc., and please do not submit composites of images related to your work. I am sure you will understand that we cannot make any promise as to whether any of your suggestions might be selected for the cover of the journal.  
+
+You can now use a single sign- on for all your accounts, view the status of all your manuscript submissions and reviews, access usage statistics for your published articles and download a record of your refereeing activity for the Nature journals.  
+
+To assist our authors in disseminating their research to the broader community, our SharedIt initiative provides you with a unique shareable link that will allow anyone (with or without a subscription) to read the published article. Recipients of the link with a subscription will also be able to download and print the PDF.
+
+<--- Page Split --->
+
+# natureresearch  
+
+You can generate the link yourself when you receive your article DOI by entering it here: <a href="http://authors.springernature.com/share">http://authors.springernature.com/share<a>.  
+
+Yours sincerely,  
+
+{REDACTED}
+
+<--- Page Split --->

peer_reviews/105ad8323cbc30432dfea0d327a986d4193c46b0744395233608b6bf22b098fc/supplementary_1_Peer Review Information./supplementary_1_Peer Review Information._det.mmd

@@ -0,0 +1,810 @@
+<|ref|>title<|/ref|><|det|>[[550, 48, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>title<|/ref|><|det|>[[116, 188, 570, 220]]<|/det|>
+# Peer Review Information  
+
+<|ref|>text<|/ref|><|det|>[[114, 245, 875, 314]]<|/det|>
+Journal: Nature Ecology & Evolution  Manuscript Title: Palaeoecological data indicates land- use changes across Europe linked to spatial heterogeneity in mortality during the Black Death pandemic  Corresponding author name(s): Alessia Masi  
+
+<|ref|>sub_title<|/ref|><|det|>[[116, 360, 568, 384]]<|/det|>
+## Reviewer Comments & Decisions:  
+
+<|ref|>title<|/ref|><|det|>[[123, 419, 353, 435]]<|/det|>
+# Decision Letter, initial version:  
+
+<|ref|>text<|/ref|><|det|>[[116, 453, 243, 468]]<|/det|>
+2nd August 2021  
+
+<|ref|>text<|/ref|><|det|>[[116, 482, 870, 513]]<|/det|>
+\*Please ensure you delete the link to your author homepage in this e- mail if you wish to forward it to your co- authors.  
+
+<|ref|>text<|/ref|><|det|>[[116, 528, 217, 543]]<|/det|>
+Dear Dr Masi,  
+
+<|ref|>text<|/ref|><|det|>[[115, 557, 879, 648]]<|/det|>
+Your manuscript entitled "Big Data Palaeoecology reveals significant variation in Black Death mortality in Europe" has now been seen by four reviewers, whose comments are attached. The reviewers have raised a number of concerns which will need to be addressed before we can offer publication in Nature Ecology & Evolution. We will therefore need to see your responses to the criticisms raised and to some editorial concerns, along with a revised manuscript, before we can reach a final decision regarding publication.  
+
+<|ref|>text<|/ref|><|det|>[[115, 662, 868, 738]]<|/det|>
+In particular, you will see that the reviewers emphasise the need to tone down the degree to which the narrative presented is a surprise from a historical perspective- - I don't think this compromises novelty since as reviewer 1 argues, it is valuable to show the palaeoecological dimension on land use change during the period. And as reviewer 4 comments, there is also a need to shore up the linkage between mortality and land use changes- - bringing more historical perspective in here could help.  
+
+<|ref|>text<|/ref|><|det|>[[116, 752, 837, 798]]<|/det|>
+We therefore invite you to revise your manuscript taking into account all reviewer and editor comments. Please highlight all changes in the manuscript text file [OPTIONAL: in Microsoft Word format].  
+
+<|ref|>text<|/ref|><|det|>[[116, 811, 876, 857]]<|/det|>
+We are committed to providing a fair and constructive peer- review process. Do not hesitate to contact us if there are specific requests from the reviewers that you believe are technically impossible or unlikely to yield a meaningful outcome.
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 50, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[116, 158, 353, 173]]<|/det|>
+When revising your manuscript:  
+
+<|ref|>text<|/ref|><|det|>[[116, 187, 845, 232]]<|/det|>
+\* Include a "Response to reviewers" document detailing, point- by- point, how you addressed each reviewer comment. If no action was taken to address a point, you must provide a compelling argument. This response will be sent back to the reviewers along with the revised manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[116, 247, 866, 291]]<|/det|>
+\* If you have not done so already please begin to revise your manuscript so that it conforms to our Article format instructions at http://www.nature.com/natecolevol/info/final- submission. Refer also to any guidelines provided in this letter.  
+
+<|ref|>text<|/ref|><|det|>[[116, 306, 855, 351]]<|/det|>
+\* Include a revised version of any required reporting checklist. It will be available to referees (and, potentially, statisticians) to aid in their evaluation if the manuscript goes back for peer review. A revised checklist is essential for re- review of the paper.  
+
+<|ref|>text<|/ref|><|det|>[[116, 366, 692, 381]]<|/det|>
+Please use the link below to submit your revised manuscript and related files:  
+
+<|ref|>sub_title<|/ref|><|det|>[[116, 396, 221, 411]]<|/det|>
+## [REDACTED]  
+
+<|ref|>text<|/ref|><|det|>[[116, 425, 866, 470]]<|/det|>
+<strong>Note: </strong> This URL links to your confidential home page and associated information about manuscripts you may have submitted, or that you are reviewing for us. If you wish to forward this email to co- authors, please delete the link to your homepage.  
+
+<|ref|>text<|/ref|><|det|>[[116, 484, 880, 529]]<|/det|>
+We hope to receive your revised manuscript within four to eight weeks. If you cannot send it within this time, please let us know. We will be happy to consider your revision so long as nothing similar has been accepted for publication at Nature Ecology & Evolution or published elsewhere.  
+
+<|ref|>text<|/ref|><|det|>[[115, 544, 872, 660]]<|/det|>
+Nature Ecology & Evolution is committed to improving transparency in authorship. As part of our efforts in this direction, we are now requesting that all authors identified as 'corresponding author' on published papers create and link their Open Researcher and Contributor Identifier (ORCID) with their account on the Manuscript Tracking System (MTS), prior to acceptance. ORCID helps the scientific community achieve unambiguous attribution of all scholarly contributions. You can create and link your ORCID from the home page of the MTS by clicking on 'Modify my Springer Nature account'. For more information please visit please visit <a href="http://www.springernature.com/orcid">www.springernature.com/orcid</a>.  
+
+<|ref|>text<|/ref|><|det|>[[115, 676, 875, 706]]<|/det|>
+Please do not hesitate to contact me if you have any questions or would like to discuss these revisions further.  
+
+<|ref|>text<|/ref|><|det|>[[115, 722, 857, 752]]<|/det|>
+We look forward to seeing the revised manuscript and thank you for the opportunity to review your work.  
+
+<|ref|>sub_title<|/ref|><|det|>[[116, 767, 221, 782]]<|/det|>
+## [REDACTED]  
+
+<|ref|>text<|/ref|><|det|>[[116, 813, 262, 827]]<|/det|>
+Reviewer expertise:  
+
+<|ref|>text<|/ref|><|det|>[[116, 842, 420, 857]]<|/det|>
+Reviewer #1: medieval economic history
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 48, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[116, 157, 407, 172]]<|/det|>
+Reviewer #2: Black Death archaeology  
+
+<|ref|>text<|/ref|><|det|>[[116, 187, 446, 201]]<|/det|>
+Reviewer #3: palaeoecology and palynology  
+
+<|ref|>text<|/ref|><|det|>[[116, 216, 446, 231]]<|/det|>
+Reviewer #4: palaeoecology and palynology  
+
+<|ref|>text<|/ref|><|det|>[[116, 261, 283, 276]]<|/det|>
+Reviewers' comments:  
+
+<|ref|>text<|/ref|><|det|>[[116, 291, 405, 306]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[116, 320, 879, 380]]<|/det|>
+This is an interesting article about the regional variation of agrarian changes in Europe during the late middle ages. Probably for the first time such a large amount of data based on pollen analysis has been brought together to study the regional variation and changes of rural cultivation activity during that period.  
+
+<|ref|>text<|/ref|><|det|>[[116, 395, 874, 455]]<|/det|>
+It confirms the existing knowledge (already since Biraben's publication of 1975 and many others after that) that the influence of the late medieval crisis hit Europe with a huge regional variation. However one of the most important merits of the article is that it is the first time that a study like this is done on the basis a large scale dataset out of the positive sciences.  
+
+<|ref|>text<|/ref|><|det|>[[115, 470, 865, 619]]<|/det|>
+However, contrary to what the title and some parts of the article are suggesting, the data do not exclusively reveal the impact of the pest epidemic of 1347- 1352. They are on the contrary revealing also other changes in the late medieval society that are due to other factors as well. Indeed other demographic catastrophes were typical for the (later) middle ages as well and certainly also are reflected in pollen analyses. We think about famines (e.g. a° 1315- 1317), wars (that became much more devastating in the late middle ages and also caused epidemic diseases) but especially other epidemics and most importantly the s.c. 'echo- (pest- )epidemics' after the first plague epidemic (e.g. 1357- 9, 1373, 1400- 1401...). However also these epidemics occurred with huge regional variations. Moreover during that period structural changes of the rural society took place that were only partly linked to the black death (e.g. land concentration and the occurrence of new elites).  
+
+<|ref|>text<|/ref|><|det|>[[115, 634, 865, 664]]<|/det|>
+To conclude: Interesting article but before publication we advise that the content should be changed as mentioned and that a new title should be formulated that is more conform to its content.  
+
+<|ref|>text<|/ref|><|det|>[[116, 708, 404, 723]]<|/det|>
+Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[116, 738, 869, 783]]<|/det|>
+This interesting study uses changes in the type of pollen in sediment cores to investigate the consequences of the black death upon the types of plant growing. The study uses a logical approach, and the illustrations explain the data convincingly.  
+
+<|ref|>text<|/ref|><|det|>[[116, 798, 877, 857]]<|/det|>
+Estimating Black Death Mortality section (lines 131- 165). You write that until now texts have been the best way to study the impact of the Black Death on population size, although it is good that you do mention papers about changing pollution levels. However, other techniques are available. One study looked at the number of fragments of pottery in test pits to show that there was a significant drop in
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 48, 872, 87]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[115, 141, 874, 207]]<|/det|>
+pottery use and production in England following the Black Death, likely a result of the smaller population size. I would encourage you to read this and give a couple of sentences on its approach (it is not my paper): Lewis, C. (2016) Disaster recovery: new archaeological evidence for the long- term impact of the 'calamitous' fourteenth century. Antiquity 90(351): 777- 797.  
+
+<|ref|>text<|/ref|><|det|>[[115, 230, 530, 261]]<|/det|>
+Minor point: Abstract first sentence: 'reknown' should read 'renown'.  
+
+<|ref|>text<|/ref|><|det|>[[116, 305, 404, 321]]<|/det|>
+Reviewer #3 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[116, 336, 864, 397]]<|/det|>
+This is a very interesting and topical paper. In principle, I was surprised by the methodological approach, but after studying it repeatedly, I find that the idea is provocative and has great explanatory power. The conclusion about the spatial and social heterogeneity of the epidemic is common sense regardless of conventional historicist approaches based on inevitably biased sources.  
+
+<|ref|>text<|/ref|><|det|>[[116, 410, 874, 471]]<|/det|>
+Obviously, pollen analysis also entails its spatial sampling, analysis and interpretation biases, in addition to the biases derived from the resolution in the identification of palynological types. Not even close to all the pollen analyses published in the studied territory are presented here either. Naturally, all these limitations are unapproachable in the current state of collaboration with Big Data.  
+
+<|ref|>text<|/ref|><|det|>[[116, 484, 868, 545]]<|/det|>
+That said, the epistemological and methodological approach of this study seems fresh and healthy to me as a pioneering study on the frontiers of knowledge. A much needed consilience between the natural sciences and the social sciences. A brilliant work in my opinion, which will be a must- read for the general public and which I predict a considerable media success.  
+
+<|ref|>text<|/ref|><|det|>[[116, 559, 868, 605]]<|/det|>
+My enthusiasm cannot but lead to the belief that the paper can be published in its current form. The language is accessible, the information well collected and the message clear and direct, without frills. Congratulations to the authors.  
+
+<|ref|>text<|/ref|><|det|>[[116, 664, 404, 679]]<|/det|>
+Reviewer #4 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 693, 860, 800]]<|/det|>
+In "Big Data Palaeoecology reveals significant variation in Black Death mortality in Europe" the authors evaluate palaeorecord of series of sedimentary sequences from 19 European countries to evaluate whether landscape changes could be linked with the hypothesis that half of the population died within a single year. This is a very interesting paper with an interesting approach and I have enjoyed reading it. It is really exciting seeing how fossil pollen time- series are answering big and relevant questions of concern to society as pandemics are. However, I have some concerns I would like the authors to consider.  
+
+<|ref|>text<|/ref|><|det|>[[116, 814, 208, 829]]<|/det|>
+Main points:  
+
+<|ref|>text<|/ref|><|det|>[[115, 842, 860, 858]]<|/det|>
+1) Mortality and landscape change: After reading the paper with interest and SI (lines 445 474) the
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 50, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[115, 141, 875, 232]]<|/det|>
+"direct" link between population decline or mortality and landscape change that I think is the main hypothesis of the paper is not clear to me. The authors used \(>1600\) palynological time- series to evaluate the demographic impact of the Black Death on a regional scale across Europe. I think that the results of the paper are not strictly quantifying mortality (e.g. what magnitude?) but responses of local/regional vegetation/landscapes to changes on land- use, and other factors including mortality. However, I think the results are still relevant.  
+
+<|ref|>text<|/ref|><|det|>[[115, 247, 880, 322]]<|/det|>
+I also found that the manuscript requires an improved discussion on demographic changes- how the authors interpret demographic changes in their results? It is mortality quantified "only" with th increase and decrease of cereals, etc? In Lines 191- 198 there are more discussion on validation. And I thank the authors for including this but I keep struggling in understanding what is the actual quantification of mortality. Also, why Poland experienced growth? More is required in here.  
+
+<|ref|>text<|/ref|><|det|>[[115, 336, 870, 382]]<|/det|>
+2) Time period: It requires more detail on why the analysis are focusing on the period between 1250 and 1450 CE. I think this detail should be explored in-depth to better discuss the otherwise relevant results.  
+
+<|ref|>text<|/ref|><|det|>[[115, 396, 881, 457]]<|/det|>
+3) Regional selection: Another issues I think it requires more description/discussion is why the authors included and excluded sites (Fig 1). I understand the limitation on finding good quality datasets for the time period of interest but I think that comparing sites located in mountain areas vs coastal areas and/or with different population densities, etc. probably impacted the findings.  
+
+<|ref|>text<|/ref|><|det|>[[115, 472, 208, 486]]<|/det|>
+Minor points  
+
+<|ref|>text<|/ref|><|det|>[[115, 500, 800, 516]]<|/det|>
+Line 113: I think it is missing the years that the pandemic was active. Please add the years.  
+
+<|ref|>text<|/ref|><|det|>[[115, 530, 861, 606]]<|/det|>
+Ecological indices: It is missing a line with the relevance of these indices to help linking landscape change and the scale of the Black Death mortality. Why are these indices important to answer the question? I read the detail in SI but still not clear to me why the authors are using them. Not suggesting that the approach is not relevant but there are so many analysis to quantify how cereals and different vegetation types change over time that it require a better justification.  
+
+<|ref|>text<|/ref|><|det|>[[115, 619, 864, 680]]<|/det|>
+Lines 171- 172: The authors suggested that "Pollen data can be used to assess past demographic dynamics as human pressure on the landscape in the preindustrial period was directly dependent on the availability of rural labour." I think this line is a bit confusing as it is mixing up to rural vs city demographics with the main point of the paper that is mortality.  
+
+<|ref|>text<|/ref|><|det|>[[115, 694, 880, 769]]<|/det|>
+Line 261: This is a very strong sentence "The significant variability of Black Death mortality that our BDP approach identifies remains..." I'm struggling to accept that the results are showing differences in mortality rates. I think that it is an interesting hypothesis but it could well be that the landscape change found may respond to differences in vegetation resilience, different baseline population densities and cultures, land- use, culture differences, etc.  
+
+<|ref|>text<|/ref|><|det|>[[115, 784, 848, 814]]<|/det|>
+Lines 286- 295 probably could be placed at the beginning of the manuscript to help integrating the spatial distribution of the Black Death and to better justify the time- period.  
+
+<|ref|>text<|/ref|><|det|>[[115, 828, 841, 858]]<|/det|>
+Lines 487- 494: Depth age models- have the authors re- calibrated them using the new calibration curves?
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 48, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[115, 157, 565, 173]]<|/det|>
+I hope my comments are helpful to improve the manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[115, 231, 523, 245]]<|/det|>
+\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*/\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*END\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*/\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\\*\*  
+
+<|ref|>title<|/ref|><|det|>[[120, 268, 395, 283]]<|/det|>
+# Author Rebuttal to Initial comments  
+
+<|ref|>sub_title<|/ref|><|det|>[[115, 305, 333, 324]]<|/det|>
+## Response to reviewers  
+
+<|ref|>sub_title<|/ref|><|det|>[[115, 370, 864, 414]]<|/det|>
+## Big Data Palaeoecology reveals significant variation in Black Death mortality in Europe  
+
+<|ref|>text<|/ref|><|det|>[[115, 426, 226, 443]]<|/det|>
+Izdebski et al.  
+
+<|ref|>text<|/ref|><|det|>[[115, 485, 863, 535]]<|/det|>
+Editor: In particular, you will see that the reviewers emphasise the need to tone down the degree to which the narrative presented is a surprise from a historical perspective- - I don't think this compromises novelty since as reviewer 1 argues, it is valuable to show the palaeoecological dimension on land use change during the period.  
+
+<|ref|>text<|/ref|><|det|>[[115, 572, 878, 648]]<|/det|>
+- > We toned down the first section of the article and took into account the non-Anglophone historiography, which contrary to the recent English-speaking consensus on the Black Death as a universal catastrophic event, noticed some regional variation during the initial phases of the second plague pandemic.  
+
+<|ref|>text<|/ref|><|det|>[[115, 684, 880, 717]]<|/det|>
+And as reviewer 4 comments, there is also a need to shore up the linkage between mortality and land use changes- - bringing more historical perspective in here could help.  
+
+<|ref|>text<|/ref|><|det|>[[115, 756, 876, 831]]<|/det|>
+- > We rewrote the relevant (first) section of the Methods, providing it with additional references. In particular, we divided this section into two clearly distinguished main points: (1) the close connection between landscape changes as revealed by palynology and population dynamics as revealed by historical sources, as proved by several recent historical-palynological studies; (2)
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 48, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[115, 142, 883, 180]]<|/det|>
+the landscape changes to be expected in plague- struck areas during the Black Death, based on the current consensus on estimated population decline among the historians of the Black Death.  
+
+<|ref|>text<|/ref|><|det|>[[115, 220, 248, 235]]<|/det|>
+Reviewer expertise:  
+
+<|ref|>text<|/ref|><|det|>[[115, 276, 386, 291]]<|/det|>
+Reviewer #1: medieval economic history  
+
+<|ref|>text<|/ref|><|det|>[[115, 331, 371, 346]]<|/det|>
+Reviewer #2: Black Death archaeology  
+
+<|ref|>text<|/ref|><|det|>[[115, 386, 409, 401]]<|/det|>
+Reviewer #3: palaeoecology and palynology  
+
+<|ref|>text<|/ref|><|det|>[[115, 441, 409, 456]]<|/det|>
+Reviewer #4: palaeoecology and palynology  
+
+<|ref|>text<|/ref|><|det|>[[115, 527, 419, 544]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 563, 861, 637]]<|/det|>
+This is an interesting article about the regional variation of agrarian changes in Europe during the late middle ages. Probably for the first time such a large amount of data based on pollen analysis has been brought together to study the regional variation and changes of rural cultivation activity during that period.  
+
+<|ref|>text<|/ref|><|det|>[[115, 656, 881, 730]]<|/det|>
+It confirms the existing knowledge (already since Biraben's publication of 1975 and many others after that) that the influence of the late medieval crisis hit Europe with a huge regional variation. However one of the most important merits of the article is that it is the first time that a study like this is done on the basis a large scale dataset out of the positive sciences.  
+
+<|ref|>text<|/ref|><|det|>[[115, 749, 880, 858]]<|/det|>
+However, contrary to what the title and some parts of the article are suggesting, the data do not exclusively reveal the impact of the pest epidemic of 1347- 1352. They are on the contrary revealing also other changes in the late medieval society that are due to other factors as well. Indeed other demographic catastrophes were typical for the (later) middle ages as well and certainly also are reflected in pollen analyses. We think about famines (e.g. a° 1315- 1317), wars (that became much more devastating in the late middle ages and also caused epidemic
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 50, 872, 87]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[115, 143, 874, 237]]<|/det|>
+diseases) but especially other epidemics and most importantly the s.c. 'echo- (pest- )epidemics' after the first plague epidemic (e.g. 1357- 9, 1373, 1400- 1401...). However also these epidemics occurred with huge regional variations. Moreover during that period structural changes of the rural society took place that were only partly linked to the black death (e.g. land concentration and the occurrence of new elites).  
+
+<|ref|>text<|/ref|><|det|>[[114, 274, 880, 571]]<|/det|>
+- > We are grateful for these careful and constructive comments. We had long discussions among the historians who co-authored this paper about the extent to which our results are connected with phenomena other than the Black Death itself. Our conclusion is that our method produces results that actually are connected directly to the Black Death understood as the initial introduction of Y. pestis to Europe during the second pandemic. Of course, the later echo-epidemics (later outbreaks) amplified the consequences of the Black Death, but the role of the outbreak of the Black Death was crucial for whether a population collapse - leading to significant landscape changes - occurred in a particular region or not. It is the Black Death itself, the pandemic of the mid-fourteenth-century, that is attributed with killing half of the population of Europe in a few years, not the subsequent outbreaks, which only added, albeit in a regionalized way, to the mortality of the Black Death decades after the Black Death. As such, it is the Black Death that we focus on. It was the major demographic moment of change in fourteenth-century European history and, in fact, in late medieval history more generally. Very few historians would expect the landscape of a region to have been altered significantly if it was not struck by the Black Death, but only by secondary and tertiary outbreaks of plague after the Black Death in the fourteenth century.  
+
+<|ref|>text<|/ref|><|det|>[[114, 580, 884, 765]]<|/det|>
+To these remark, we should add that our method is not focused on showing the different trajectories of the late medieval crisis (as was, for instance, Izdebski et al. 2016, https://doi.org/10.1016/j.eeh.2015.10.003); instead of focusing on time series and trend analysis, we focus on the extent to which a single event, well- constrained in time - the Black Death - can be associated with landscape changes across Europe. It may well be that all the other phenomena that are collectively known as the late medieval crisis left their imprint on the landscape as well (for instance, our data show the impact of the Great Famine of 1315- 1317 in Devon/England, for which we have no significant impact of the Black Death), but our method in this particular article is set up in such a way that it does not reveal them: it is very clearly focused on the Black Death.  
+
+<|ref|>text<|/ref|><|det|>[[115, 775, 857, 849]]<|/det|>
+Hence, the results presented in our article concern 100- year periods before and after the Black Death, that is 4 human generations (the minimum time to overcome a sudden population loss in the range of 30- 50%). Additionally, in the Supplementary Material, we already showed the results for 50 and 25 year periods before and after the Black Death: they
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 50, 872, 87]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[114, 143, 878, 403]]<|/det|>
+are very similar to the results of the 100 year periods, as we discussed in the main text. To emphasise this fact, and further demonstrate the focus of our analysis on the Black Death in particular, we now added additional figures to the text and commented on them. Based on the consistency between these different periods of analysis, we are convinced that the landscape changes we analyse across Europe are related directly to the Black Death, and not - as the reviewer suggests - to the 'Dantean Anomaly' and the Great Famine of the 1310s or to the late fourteenth or early fifteenth century- plague outbreaks. The impact of these phenomena is not captured by the 25 year or even 50 year periods of analysis). Thus, given the current consensus hypothesis we want to test (universal mass mortality), we are able to demonstrate that whereas it did occur in some areas (e.g. Greece, Sweden or Italy), it did not occur elsewhere (e.g. Poland or Spain). We believe this is the main merit of our article, and to some extent it is indeed confirmation of historical works of Biraben, Malwost or Gottfried, who emphasised regional differentiation of the late medieval crisis in Europe, a minority perspective today which we now make clear reference to in the article).  
+
+<|ref|>text<|/ref|><|det|>[[116, 420, 857, 476]]<|/det|>
+To conclude: Interesting article but before publication we advise that the content should be changed as mentioned and that a new title should be formulated that is more conform to its content.  
+
+<|ref|>text<|/ref|><|det|>[[116, 494, 866, 531]]<|/det|>
+- > We have suggested a new title, clearly focused on the vision of the Black Death that dominates plague studies and the public imagination - the vision which we falsify in our study.  
+
+<|ref|>text<|/ref|><|det|>[[116, 550, 418, 567]]<|/det|>
+Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[116, 586, 855, 642]]<|/det|>
+This interesting study uses changes in the type of pollen in sediment cores to investigate the consequences of the black death upon the types of plant growing. The study uses a logical approach, and the illustrations explain the data convincingly.  
+
+<|ref|>text<|/ref|><|det|>[[114, 660, 883, 825]]<|/det|>
+Estimating Black Death Mortality section (lines 131- 165). You write that until now texts have been the best way to study the impact of the Black Death on population size, although it is good that you do mention papers about changing pollution levels. However, other techniques are available. One study looked at the number of fragments of pottery in test pits to show that there was a significant drop in pottery use and production in England following the Black Death, likely a result of the smaller population size. I would encourage you to read this and give a couple of sentences on its approach (it is not my paper): Lewis, C. (2016) Disaster recovery: new archaeological evidence for the long- term impact of the 'calamitous' fourteenth century. Antiquity 90(351): 777- 797.
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 48, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[115, 144, 844, 200]]<|/det|>
+- > We have now included this study in our method section. We also would like to thank the reviewer 2 for the positive comments on our paper and for expanding even further our interdisciplinarity.  
+
+<|ref|>text<|/ref|><|det|>[[115, 228, 566, 264]]<|/det|>
+Minor point: Abstract first sentence: 'reknown' should read 'renown'.  
+
+<|ref|>text<|/ref|><|det|>[[116, 293, 309, 310]]<|/det|>
+- > Corrected, thank you.  
+
+<|ref|>text<|/ref|><|det|>[[115, 348, 418, 366]]<|/det|>
+Reviewer #3 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 385, 882, 477]]<|/det|>
+This is a very interesting and topical paper. In principle, I was surprised by the methodological approach, but after studying it repeatedly, I find that the idea is provocative and has great explanatory power. The conclusion about the spatial and social heterogeneity of the epidemic is common sense regardless of conventional historicist approaches based on inevitably biased sources.  
+
+<|ref|>text<|/ref|><|det|>[[115, 496, 874, 589]]<|/det|>
+Obviously, pollen analysis also entails its spatial sampling, analysis and interpretation biases, in addition to the biases derived from the resolution in the identification of palynological types. Not even close to all the pollen analyses published in the studied territory are presented here either. Naturally, all these limitations are unapproachable in the current state of collaboration with Big Data.  
+
+<|ref|>text<|/ref|><|det|>[[115, 607, 872, 682]]<|/det|>
+That said, the epistemological and methodological approach of this study seems fresh and healthy to me as a pioneering study on the frontiers of knowledge. A much needed consilience between the natural sciences and the social sciences. A brilliant work in my opinion, which will be a must- read for the general public and which I predict a considerable media success.  
+
+<|ref|>text<|/ref|><|det|>[[115, 700, 840, 755]]<|/det|>
+My enthusiasm cannot but lead to the belief that the paper can be published in its current form. The language is accessible, the information well collected and the message clear and direct, without frills. Congratulations to the authors.  
+
+<|ref|>text<|/ref|><|det|>[[115, 774, 568, 792]]<|/det|>
+- > Thank you very much for such an encouraging review!  
+
+<|ref|>text<|/ref|><|det|>[[115, 830, 418, 848]]<|/det|>
+Reviewer #4 (Remarks to the Author):
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 50, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[115, 161, 870, 292]]<|/det|>
+In "Big Data Palaeoecology reveals significant variation in Black Death mortality in Europe" the authors evaluate paleorecord of series of sedimentary sequences from 19 European countries to evaluate whether landscape changes could be linked with the hypothesis that half of the population died within a single year. This is a very interesting paper with an interesting approach and I have enjoyed reading it. It is really exciting seeing how fossil pollen time- series are answering big and relevant questions of concern to society as pandemics are. However, I have some concerns I would like the authors to consider.  
+
+<|ref|>text<|/ref|><|det|>[[115, 312, 217, 328]]<|/det|>
+Main points:  
+
+<|ref|>text<|/ref|><|det|>[[115, 346, 881, 476]]<|/det|>
+1) Mortality and landscape change: After reading the paper with interest and SI (lines 445 474) the "direct" link between population decline or mortality and landscape change that I think is the main hypothesis of the paper is not clear to me. The authors used >1600 palynological time-series to evaluate the demographic impact of the Black Death on a regional scale across Europe. I think that the results of the paper are not strictly quantifying mortality (e.g. what magnitude?) but responses of local/regional vegetation/landscapes to changes on land-use, and other factors including mortality. However, I think the results are still relevant.  
+
+<|ref|>text<|/ref|><|det|>[[115, 515, 876, 625]]<|/det|>
+-> Yes, indeed. In agreement with the comments of R1, and by changing the title, we now tried to make clear that while we do study landscape changes (or lack thereof), we do so entirely in order to test the currently dominant view that the Black Death led to universal mortality within the range of 30-50%. We now make clear in the methods section why this is important to us: such a truly significant mortality would lead to profound and sudden landscape change, which we indeed observe in some of our study regions.  
+
+<|ref|>text<|/ref|><|det|>[[115, 644, 878, 737]]<|/det|>
+I also found that the manuscript requires an improved discussion on demographic changes- how the authors interpret demographic changes in their results? It is mortality quantified "only" with th increase and decrease of cereals, etc? In Lines 191- 198 there are more discussion on validation. And I thank the authors for including this but I keep struggling in understanding what is the actual quantification of mortality.  
+
+<|ref|>text<|/ref|><|det|>[[115, 776, 860, 849]]<|/det|>
+-> Entirely understood. We have now improved our discussion of the connection to plague mortality in the relevant sections of the Methods section and in the main text. As we explain, we do not provide direct, concrete estimates of mortality, as the state of the research on palynology and historical demography has not yet reached this stage, despite some major
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 49, 872, 87]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[114, 143, 880, 291]]<|/det|>
+recent advances in this respect. Our aim instead is to use pollen data to verify the popular vision of the Black Death as a universal mortality catastrophe in Europe, leading to \(30 - 50\%\) of the population dying in a few mere years. As we now try to explain more clearly, such a scenario would result in major landscape changes following the sudden and profound reduction of human pressure on the landscape. We observe significant population erosion in some parts of Europe (consistently with the limited written evidence available for Black Death mortality), while not in others (for which we provide explanation in the concluding section of the main text).  
+
+<|ref|>text<|/ref|><|det|>[[116, 330, 628, 348]]<|/det|>
+Also, why Poland experienced growth? More is required in here.  
+
+<|ref|>text<|/ref|><|det|>[[117, 387, 337, 404]]<|/det|>
+- > We now explain this fact.  
+
+<|ref|>text<|/ref|><|det|>[[116, 424, 872, 478]]<|/det|>
+2) Time period: It requires more detail on why the analysis are focusing on the period between 1250 and 1450 CE. I think this detail should be explored in-depth to better discuss the otherwise relevant results.  
+
+<|ref|>text<|/ref|><|det|>[[116, 517, 833, 572]]<|/det|>
+- > We now discuss our choice of 100 year slices of time on either side of 1350 as our main period of analysis in the first section of the Methods. We now also discuss the results we obtained for shorter time periods in the main text.  
+
+<|ref|>text<|/ref|><|det|>[[115, 591, 870, 683]]<|/det|>
+3) Regional selection: Another issues I think it requires more description/discussion is why the authors included and excluded sites (Fig 1). I understand the limitation on finding good quality datasets for the time period of interest but I think that comparing sites located in mountain areas vs coastal areas and/or with different population densities, etc. probably impacted the findings.  
+
+<|ref|>text<|/ref|><|det|>[[115, 712, 881, 842]]<|/det|>
+- > We now provide two additional supplementary figures that show the geographical setting (lowland vs highland) of our sites as well as the population densities around our sites in the fourteenth century. These variables are not linked to the results we obtained for our BDP pollen indicators, however. Please note that we undertook as well a comparative study of change in the landscape over periods of 200, 100 and 50 years. We assume that over such short time scales physical geographical characteristics of each site remain stable and do not impact our results and conclusions.
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 50, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[115, 202, 220, 218]]<|/det|>
+Minor points  
+
+<|ref|>text<|/ref|><|det|>[[115, 238, 835, 256]]<|/det|>
+Line 113: I think it is missing the years that the pandemic was active. Please add the years.  
+
+<|ref|>text<|/ref|><|det|>[[115, 285, 872, 378]]<|/det|>
+- > This Black Death pandemic probably began in the 1330s in Central Asia and in many parts of Eurasia and Africa - outside of Europe - continued until the late 1350s. For this reason, we believe providing an approximate chronological range of the "mid-14th c." is most correct. For the entire Eurasia and Africa, providing set years would be difficult to substantiate and there is no consensus (contrary to the dates for Europe only).  
+
+<|ref|>text<|/ref|><|det|>[[115, 406, 881, 516]]<|/det|>
+Ecological indices: It is missing a line with the relevance of these indices to help linking landscape change and the scale of the Black Death mortality. Why are these indices important to answer the question? I read the detail in SI but still not clear to me why the authors are using them. Not suggesting that the approach is not relevant but there are so many analysis to quantify how cereals and different vegetation types change over time that it require a better justification.  
+
+<|ref|>text<|/ref|><|det|>[[115, 556, 761, 573]]<|/det|>
+- > We have now provided more explanation in the main text and in the Methods.  
+
+<|ref|>text<|/ref|><|det|>[[115, 592, 881, 666]]<|/det|>
+Lines 171- 172: The authors suggested that "Pollen data can be used to assess past demographic dynamics as human pressure on the landscape in the preindustrial period was directly dependent on the availability of rural labour." I think this line is a bit confusing as it is mixing up to rural vs city demographics with the main point of the paper that is mortality.  
+
+<|ref|>text<|/ref|><|det|>[[115, 705, 857, 760]]<|/det|>
+- > As we explain in the Methods, \(\sim 90\%\) of the late medieval European population lived in the countryside, so in order to understand overall mortality levels during the Black Death, our primary focus should be the rural populations.  
+
+<|ref|>text<|/ref|><|det|>[[115, 789, 877, 844]]<|/det|>
+Line 261: This is a very strong sentence "The significant variability of Black Death mortality that our BDP approach identifies remains..." I'm struggling to accept that the results are showing differences in mortality rates. I think that it is an interesting hypothesis but it could well be that
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 50, 872, 87]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[115, 143, 836, 180]]<|/det|>
+the landscape change found may respond to differences in vegetation resilience, different baseline population densities and cultures, land- use, culture differences, etc.  
+
+<|ref|>text<|/ref|><|det|>[[114, 209, 872, 412]]<|/det|>
+- > Duly noted. As we argue in the Methods, though, the changes we observe are related to human pressure on the landscape, which in turn was dependent on population levels. We also focus primarily on cultivated plants, cereals, whose presence in the landscape diminishes dramatically and over very short time scales without human activity/pressure that supports these plants, as we explain in the Methods. Moreover, thanks to our strict chronological focus, also as now explained further in response to the comments of R1, we are in the position to argue that these landscape changes (or lack thereof) were directly related to the Black Death, that is, the commonly supposed most important demographic event of the fourteenth century. At the same time, we agree with R4 that the factors s/he mentions were important as moderating the Black Death's impact at the regional and local level - and we discuss this point in the concluding section of the main text.  
+
+<|ref|>text<|/ref|><|det|>[[115, 440, 867, 478]]<|/det|>
+Lines 286- 295 probably could be placed at the beginning of the manuscript to help integrating the spatial distribution of the Black Death and to better justify the time- period.  
+
+<|ref|>text<|/ref|><|det|>[[115, 506, 800, 543]]<|/det|>
+- > We now focus more on explaining the temporal and spatial focus we have adopted throughout the text.  
+
+<|ref|>text<|/ref|><|det|>[[115, 572, 800, 608]]<|/det|>
+Lines 487- 494: Depth age models- have the authors re- calibrated them using the new calibration curves?  
+
+<|ref|>text<|/ref|><|det|>[[115, 648, 875, 684]]<|/det|>
+- > Yes, absolutely: as we explain in the Methods, we re- calculated age-depth models for a large portion of our sites: especially for those that were using the older calibration curves.  
+
+<|ref|>text<|/ref|><|det|>[[115, 704, 597, 722]]<|/det|>
+I hope my comments are helpful to improve the manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[116, 752, 412, 769]]<|/det|>
+- > Yes, indeed, thank you very much!
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[549, 48, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>sub_title<|/ref|><|det|>[[123, 181, 345, 197]]<|/det|>
+## Decision Letter, first revision:  
+
+<|ref|>text<|/ref|><|det|>[[120, 216, 250, 231]]<|/det|>
+8th October 2021  
+
+<|ref|>text<|/ref|><|det|>[[116, 245, 223, 260]]<|/det|>
+Dear Dr. Masi,  
+
+<|ref|>text<|/ref|><|det|>[[115, 274, 879, 425]]<|/det|>
+Thank you for submitting your revised manuscript "The Black Death Killed Far Fewer Europeans than Commonly Thought, Big Data Paleoecology Reveals" (NATECoLEVEL- 210614034A). It has now been seen again by the original reviewers and their comments are below. The reviewers find that the paper has improved in revision, and therefore we'll be happy in principle to publish it in Nature Ecology & Evolution, pending revisions to satisfy the reviewers' final requests and to comply with our editorial and formatting guidelines. As explained below, our full recommendations will be sent in a later email, but for now, please note that both reviewers recommend that considerable nuance be brought into the discussion--your main text is on the short side by my rough calculation, so I estimate that you should be able to incorporate up to about 500 words of additional discussion which should address the remaining reviewer concerns.  
+
+<|ref|>text<|/ref|><|det|>[[115, 438, 857, 470]]<|/det|>
+If the current version of your manuscript is in a PDF format, please email us a copy of the file in an editable format (Microsoft Word or LaTex)-- we can not proceed with PDFs at this stage.  
+
+<|ref|>text<|/ref|><|det|>[[116, 483, 860, 529]]<|/det|>
+We are now performing detailed checks on your paper and will send you a checklist detailing our editorial and formatting requirements in about a week. Please do not upload the final materials and make any revisions until you receive this additional information from us.  
+
+<|ref|>text<|/ref|><|det|>[[115, 543, 870, 574]]<|/det|>
+Thank you again for your interest in Nature Ecology & Evolution. Please do not hesitate to contact me if you have any questions.  
+
+<|ref|>sub_title<|/ref|><|det|>[[116, 587, 221, 603]]<|/det|>
+## [REDACTED]  
+
+<|ref|>text<|/ref|><|det|>[[116, 618, 404, 633]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>sub_title<|/ref|><|det|>[[116, 648, 273, 662]]<|/det|>
+## GENERAL COMMENT:  
+
+<|ref|>text<|/ref|><|det|>[[115, 663, 877, 693]]<|/det|>
+The article has been upgraded a lot: the amount of relevant literature has been increased and the text is more nuanced.  
+
+<|ref|>text<|/ref|><|det|>[[115, 694, 564, 709]]<|/det|>
+A few remarks anyway (put in the reply- text of the authors)  
+
+<|ref|>sub_title<|/ref|><|det|>[[115, 723, 188, 737]]<|/det|>
+## DETAILS:  
+
+<|ref|>text<|/ref|><|det|>[[115, 739, 457, 753]]<|/det|>
+Reply text of the authors to the first remarks:  
+
+<|ref|>text<|/ref|><|det|>[[115, 753, 876, 797]]<|/det|>
+"Very few historians would expect the landscape of a region to have been altered significantly if it was not struck by the Black Death, but only by secondary and tertiary outbreaks of plague after the Black Death in the fourteenth century."  
+
+<|ref|>text<|/ref|><|det|>[[115, 798, 585, 858]]<|/det|>
+Reply of the reviewer: Anyway there were important regions where the later outbreaks had much more influence that the first outbreak. Moreover if there would not have been 'echo- epidemics', the economy (and the landscapes) probably would
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 48, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[116, 143, 277, 157]]<|/det|>
+have recovered early.  
+
+<|ref|>text<|/ref|><|det|>[[116, 173, 405, 187]]<|/det|>
+Reply text of the authors to the review  
+
+<|ref|>text<|/ref|><|det|>[[115, 188, 879, 321]]<|/det|>
+"To these remark, we should add that our method is not focused on showing the different trajectories of the late medieval crisis (as was, for instance, Izdebski et al. 2016, https://doi.org/10.1016/j.eeh.2015.10.003); instead of focusing on time series and trend analysis, we focus on the extent to which a single event, well constrained in time - the Black Death - can be associated with landscape changes across Europe. It may well be that all the other phenomena that are collectively known as the late medieval crisis left their imprint on the landscape as well (for instance, our data show the impact of the Great Famine of 1315- 1317 in Devon/England, for which we have no significant impact of the Black Death), but our method in this particular article is set up in such a way that it does not reveal them: it is very clearly focused on the Black Death."  
+
+<|ref|>text<|/ref|><|det|>[[116, 337, 280, 351]]<|/det|>
+Reply of the reviewer:  
+
+<|ref|>text<|/ref|><|det|>[[116, 352, 560, 425]]<|/det|>
+My remark remains...Why should the data reveal more the influence of the black dead than other elements of the rural economy that were changing in the later middle ages? Of course it is logic that one cannot deal with these other elements but at least it could be mentioned.  
+
+<|ref|>text<|/ref|><|det|>[[116, 470, 404, 485]]<|/det|>
+Reviewer #4 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 500, 880, 604]]<|/det|>
+This version of the manuscript reads very well and I thank the authors for taking into consideration my comments. This paper is interesting, well written, with fantastic figures, and can bring good discussion on how to integrate data from different paleo- sciences. I agree that population changes may result in major landscape changes (accounted using pollen) following the sudden and profound reduction of human pressure on the landscape that the BD produced- we have a good example with COVID- 19. However, I keep struggling with the concept that the land- use changes found are reflecting differences in mortality rates or the magnitude of the severity of BD, only (line 313, title, etc).  
+
+<|ref|>text<|/ref|><|det|>[[115, 619, 882, 724]]<|/det|>
+Overall, the authors should tone down a bit more the narrative around novelty (e.g. pioneering). Using pollen data (or Big- data palaeoecology as the authors refers too) to infer land- use changes or landscape changes is not novel as it is not the link between land- use change and population dynamics. What is novel and exciting of this paper is the integration of historical and paleoecological sources to answer a current important research question: what happen to the landscape when there is a global pandemic with all the associated factors (economical crisis, rates of mortality, duration of the pandemic, type of vegetation, etc...)?  
+
+<|ref|>text<|/ref|><|det|>[[116, 739, 250, 753]]<|/det|>
+Interesting paper!
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 48, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[116, 160, 255, 174]]<|/det|>
+13th October 2021  
+
+<|ref|>text<|/ref|><|det|>[[116, 203, 223, 217]]<|/det|>
+Dear Dr. Masi,  
+
+<|ref|>text<|/ref|><|det|>[[115, 232, 872, 337]]<|/det|>
+Thank you for your patience as we've prepared the guidelines for final submission of your Nature Ecology & Evolution manuscript, "The Black Death Killed Far Fewer Europeans than Commonly Thought, Big Data Paleoecology Reveals" (NATECOLEVOL- 210614034A). Please carefully follow the step- by- step instructions provided in the attached file, and add a response in each row of the table to indicate the changes that you have made. Please also check and comment on any additional marked- up edits we have proposed within the text. Ensuring that each point is addressed will help to ensure that your revised manuscript can be swiftly handed over to our production team.  
+
+<|ref|>text<|/ref|><|det|>[[116, 351, 865, 397]]<|/det|>
+\*\*We would like to start working on your revised paper, with all of the requested files and forms, as soon as possible (preferably within two weeks). Please get in contact with us immediately if you anticipate it taking more than two weeks to submit these revised files.\*\*  
+
+<|ref|>text<|/ref|><|det|>[[115, 411, 839, 441]]<|/det|>
+When you upload your final materials, please include a point- by- point response to any remaining reviewer comments.  
+
+<|ref|>text<|/ref|><|det|>[[115, 455, 868, 515]]<|/det|>
+If you have not done so already, please alert us to any related manuscripts from your group that are under consideration or in press at other journals, or are being written up for submission to other journals (see: https://www.nature.com/nature- research/editorial- policies/plagiarism#policy- on- duplicate- publication for details).  
+
+<|ref|>text<|/ref|><|det|>[[115, 529, 880, 604]]<|/det|>
+In recognition of the time and expertise our reviewers provide to Nature Ecology & Evolution's editorial process, we would like to formally acknowledge their contribution to the external peer review of your manuscript entitled "The Black Death Killed Far Fewer Europeans than Commonly Thought, Big Data Paleoecology Reveals". For those reviewers who give their assent, we will be publishing their names alongside the published article.  
+
+<|ref|>text<|/ref|><|det|>[[115, 618, 878, 724]]<|/det|>
+Nature Ecology & Evolution offers a Transparent Peer Review option for new original research manuscripts submitted after December 1st, 2019. As part of this initiative, we encourage our authors to support increased transparency into the peer review process by agreeing to have the reviewer comments, author rebuttal letters, and editorial decision letters published as a Supplementary item. When you submit your final files please clearly state in your cover letter whether or not you would like to participate in this initiative. Please note that failure to state your preference will result in delays in accepting your manuscript for publication.  
+
+<|ref|>text<|/ref|><|det|>[[118, 739, 324, 753]]<|/det|>
+<b>Cover suggestions</b>  
+
+<|ref|>text<|/ref|><|det|>[[115, 768, 816, 798]]<|/det|>
+As you prepare your final files we encourage you to consider whether you have any images or illustrations that may be appropriate for use on the cover of Nature Ecology & Evolution.  
+
+<|ref|>text<|/ref|><|det|>[[115, 812, 880, 857]]<|/det|>
+Covers should be both aesthetically appealing and scientifically relevant, and should be supplied at the best quality available. Due to the prominence of these images, we do not generally select images featuring faces, children, text, graphs, schematic drawings, or collages on our covers.
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 48, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[115, 157, 844, 188]]<|/det|>
+We accept TIFF, JPEG, PNG or PSD file formats (a layered PSD file would be ideal), and the image should be at least 300ppi resolution (preferably 600- 1200 ppi), in CMYK colour mode.  
+
+<|ref|>text<|/ref|><|det|>[[115, 202, 875, 232]]<|/det|>
+If your image is selected, we may also use it on the journal website as a banner image, and may need to make artistic alterations to fit our journal style.  
+
+<|ref|>text<|/ref|><|det|>[[115, 247, 860, 277]]<|/det|>
+Please submit your suggestions, clearly labeled, along with your final files. We'll be in touch if more information is needed.  
+
+<|ref|>text<|/ref|><|det|>[[115, 306, 874, 397]]<|/det|>
+Nature Ecology & Evolution has now transitioned to a unified Rights Collection system which will allow our Author Services team to quickly and easily collect the rights and permissions required to publish your work. Approximately 10 days after your paper is formally accepted, you will receive an email in providing you with a link to complete the grant of rights. If your paper is eligible for Open Access, our Author Services team will also be in touch regarding any additional information that may be required to arrange payment for your article.  
+
+<|ref|>text<|/ref|><|det|>[[115, 410, 868, 500]]<|/det|>
+Please note that <i>Nature Ecology & Evolution</i> is a Transformative Journal (TJ). Authors may publish their research with us through the traditional subscription access route or make their paper immediately open access through payment of an article- processing charge (APC). Authors will not be required to make a final decision about access to their article until it has been accepted. <a href="https://www.springernature.com/gp/open- research/transformative- journals"> Find out more about Transformative Journals</a>  
+
+<|ref|>text<|/ref|><|det|>[[115, 515, 876, 679]]<|/det|>
+<B>Authors may need to take specific actions to achieve <a href="https://www.springernature.com/gp/open- research/funding/policy- compliance- faqs"> compliance</a> with funder and institutional open access mandates.</b> For submissions from January 2021, if your research is supported by a funder that requires immediate open access (e.g. according to <a href="https://www.springernature.com/gp/open- research/plan- s- compliance">Plan S principles</a>) then you should select the gold OA route, and we will direct you to the compliant route where possible. For authors selecting the subscription publication route our standard licensing terms will need to be accepted, including our <a href="https://www.springernature.com/gp/open- research/policies/journal- policies">self- archiving policies</a>. Those standard licensing terms will supersede any other terms that the author or any third party may assert apply to any version of the manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[115, 694, 848, 724]]<|/det|>
+Please note that you will not receive your proofs until the publishing agreement has been received through our system.  
+
+<|ref|>text<|/ref|><|det|>[[115, 739, 866, 799]]<|/det|>
+For information regarding our different publishing models please see our <a href="https://www.springernature.com/gp/open- research/transformative- journals"> Transformative Journals </a> page. If you have any questions about costs, Open Access requirements, or our legal forms, please contact ASJournals@springernature.com.  
+
+<|ref|>text<|/ref|><|det|>[[115, 843, 553, 858]]<|/det|>
+Please use the following link for uploading these materials:
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 48, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>sub_title<|/ref|><|det|>[[115, 142, 222, 159]]<|/det|>
+## [REDACTED]  
+
+<|ref|>text<|/ref|><|det|>[[115, 173, 601, 189]]<|/det|>
+If you have any further questions, please feel free to contact me.  
+
+<|ref|>sub_title<|/ref|><|det|>[[115, 217, 222, 233]]<|/det|>
+## [REDACTED]  
+
+<|ref|>text<|/ref|><|det|>[[115, 262, 293, 309]]<|/det|>
+Reviewer #1: Remarks to the Author: GENERAL COMMENT:  
+
+<|ref|>text<|/ref|><|det|>[[115, 309, 880, 339]]<|/det|>
+The article has been upgraded a lot: the amount of relevant literature has been increased and the text is more nuanced.  
+
+<|ref|>text<|/ref|><|det|>[[115, 339, 564, 354]]<|/det|>
+A few remarks anyway (put in the reply- text of the authors)  
+
+<|ref|>text<|/ref|><|det|>[[115, 368, 188, 381]]<|/det|>
+DETAILS:  
+
+<|ref|>text<|/ref|><|det|>[[115, 382, 457, 396]]<|/det|>
+Reply text of the authors to the first remarks:  
+
+<|ref|>text<|/ref|><|det|>[[115, 397, 875, 440]]<|/det|>
+"Very few historians would expect the landscape of a region to have been altered significantly if it was not struck by the Black Death, but only by secondary and tertiary outbreaks of plague after the Black Death in the fourteenth century."  
+
+<|ref|>text<|/ref|><|det|>[[115, 441, 585, 516]]<|/det|>
+Reply of the reviewer: Anyway there were important regions where the later outbreaks had much more influence that the first outbreak. Moreover if there would not have been 'echoepidemics', the economy (and the landscapes) probably would have recovered early.  
+
+<|ref|>text<|/ref|><|det|>[[115, 530, 405, 545]]<|/det|>
+Reply text of the authors to the review  
+
+<|ref|>text<|/ref|><|det|>[[115, 546, 877, 679]]<|/det|>
+"To these remark, we should add that our method is not focused on showing the different trajectories of the late medieval crisis (as was, for instance, Izdebski et al. 2016, https://doi.org/10.1016/j.eeh.2015.10.003); instead of focusing on time series and trend analysis, we focus on the extent to which a single event, well constrained in time - the Black Death - can be associated with landscape changes across Europe. It may well be that all the other phenomena that are collectively known as the late medieval crisis left their imprint on the landscape as well (for instance, our data show the impact of the Great Famine of 1315- 1317 in Devon/England, for which we have no significant impact of the Black Death), but our method in this particular article is set up in such a way that it does not reveal them: it is very clearly focused on the Black Death."  
+
+<|ref|>text<|/ref|><|det|>[[115, 694, 280, 708]]<|/det|>
+Reply of the reviewer:  
+
+<|ref|>text<|/ref|><|det|>[[115, 709, 560, 783]]<|/det|>
+My remark remains...Why should the data reveal more the influence of the black dead than other elements of the rural economy that were changing in the later middle ages? Of course it is logic that one cannot deal with these other elements but at least it could be mentioned.  
+
+<|ref|>text<|/ref|><|det|>[[115, 828, 291, 858]]<|/det|>
+Reviewer #4: Remarks to the Author:
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 50, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[115, 141, 880, 248]]<|/det|>
+This version of the manuscript reads very well and I thank the authors for taking into consideration my comments. This paper is interesting, well written, with fantastic figures, and can bring good discussion on how to integrate data from different paleo- sciences. I agree that population changes may result in major landscape changes (accounted using pollen) following the sudden and profound reduction of human pressure on the landscape that the BD produced- we have a good example with COVID- 19. However, I keep struggling with the concept that the land- use changes found are reflecting differences in mortality rates or the magnitude of the severity of BD, only (line 313, title, etc).  
+
+<|ref|>text<|/ref|><|det|>[[115, 261, 881, 367]]<|/det|>
+Overall, the authors should tone down a bit more the narrative around novelty (e.g. pioneering). Using pollen data (or Big- data palaeoecology as the authors refers too) to infer land- use changes or landscape changes is not novel as it is not the link between land- use change and population dynamics. What is novel and exciting of this paper is the integration of historical and paleoecological sources to answer a current important research question: what happen to the landscape when there is a global pandemic with all the associated factors (economical crisis, rates of mortality, duration of the pandemic, type of vegetation, etc...)?  
+
+<|ref|>text<|/ref|><|det|>[[116, 381, 250, 396]]<|/det|>
+Interesting paper!  
+
+<|ref|>text<|/ref|><|det|>[[120, 415, 354, 431]]<|/det|>
+Author Rebuttal, first revision:  
+
+<|ref|>sub_title<|/ref|><|det|>[[120, 476, 805, 501]]<|/det|>
+## Response to the reviewers' comments (second revision)  
+
+<|ref|>text<|/ref|><|det|>[[115, 510, 850, 590]]<|/det|>
+Reviewer #1: Remarks to the Author: GENERAL COMMENT: The article has been upgraded a lot: the amount of relevant literature has been increased and the text is more nuanced.  
+
+<|ref|>text<|/ref|><|det|>[[115, 630, 208, 644]]<|/det|>
+- > Thank you!  
+
+<|ref|>text<|/ref|><|det|>[[115, 673, 512, 688]]<|/det|>
+A few remarks anyway (put in the reply- text of the authors)  
+
+<|ref|>text<|/ref|><|det|>[[115, 706, 175, 720]]<|/det|>
+DETAILS:  
+
+<|ref|>text<|/ref|><|det|>[[115, 723, 419, 738]]<|/det|>
+Reply text of the authors to the first remarks:  
+
+<|ref|>text<|/ref|><|det|>[[115, 740, 875, 787]]<|/det|>
+"Very few historians would expect the landscape of a region to have been altered significantly if it was not struck by the Black Death, but only by secondary and tertiary outbreaks of plague after the Black Death in the fourteenth century."  
+
+<|ref|>text<|/ref|><|det|>[[115, 789, 521, 837]]<|/det|>
+Reply of the reviewer: Anyway there were important regions where the later outbreaks had much more influence that the first outbreak. Moreover if there would not have been 'echo
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 48, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[115, 143, 536, 176]]<|/det|>
+epidemics', the economy (and the landscapes) probably would have recovered early.  
+
+<|ref|>text<|/ref|><|det|>[[116, 214, 448, 228]]<|/det|>
+- > We took account of this point in the discussion.  
+
+<|ref|>text<|/ref|><|det|>[[116, 247, 375, 262]]<|/det|>
+Reply text of the authors to the review  
+
+<|ref|>text<|/ref|><|det|>[[115, 263, 877, 396]]<|/det|>
+"To these remark, we should add that our method is not focused on showing the different trajectories of the late medieval crisis (as was, for instance, Izdebski et al. 2016, https://doi.org/10.1016/j.eeh.2015.10.003); instead of focusing on time series and trend analysis, we focus on the extent to which a single event, well constrained in time - the Black Death - can be associated with landscape changes across Europe. It may well be that all the other phenomena that are collectively known as the late medieval crisis left their imprint on the landscape as well (for instance, our data show the impact of the Great Famine of 1315- 1317 in Devon/England, for which we have no significant impact of the Black Death), but our method in this particular article is set up in such a way that it does not reveal them: it is very clearly focused on the Black Death."  
+
+<|ref|>text<|/ref|><|det|>[[115, 414, 264, 428]]<|/det|>
+Reply of the reviewer:  
+
+<|ref|>text<|/ref|><|det|>[[115, 430, 516, 512]]<|/det|>
+My remark remains...Why should the data reveal more the influence of the black dead than other elements of the rural economy that were changing in the later middle ages? Of course it is logic that one cannot deal with these other elements but at least it could be mentioned.  
+
+<|ref|>text<|/ref|><|det|>[[116, 530, 440, 545]]<|/det|>
+- > As suggested, we now discuss these elements.  
+
+<|ref|>text<|/ref|><|det|>[[115, 564, 203, 577]]<|/det|>
+Reviewer #4:  
+
+<|ref|>text<|/ref|><|det|>[[116, 581, 273, 594]]<|/det|>
+Remarks to the Author:  
+
+<|ref|>text<|/ref|><|det|>[[115, 596, 872, 713]]<|/det|>
+This version of the manuscript reads very well and I thank the authors for taking into consideration my comments. This paper is interesting, well written, with fantastic figures, and can bring good discussion on how to integrate data from different paleo- sciences. I agree that population changes may result in major landscape changes (accounted using pollen) following the sudden and profound reduction of human pressure on the landscape that the BD produced- we have a good example with COVID- 19. However, I keep struggling with the concept that the land- use changes found are reflecting differences in mortality rates or the magnitude of the severity of BD, only (line 313, title, etc).  
+
+<|ref|>text<|/ref|><|det|>[[115, 748, 871, 764]]<|/det|>
+- > We modified the text in order to account for the context-dependent complexities involved in this potential link.  
+
+<|ref|>text<|/ref|><|det|>[[115, 782, 876, 849]]<|/det|>
+Overall, the authors should tone down a bit more the narrative around novelty (e.g. pioneering). Using pollen data (or Big- data palaeoecology as the authors refers too) to infer land- use changes or landscape changes is not novel as it is not the link between land- use change and population dynamics. What is novel and exciting of this paper is the integration of historical and paleoecological sources to answer a current important research question: what
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 50, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[115, 143, 882, 177]]<|/det|>
+happen to the landscape when there is a global pandemic with all the associated factors (economical crisis, rates of mortality, duration of the pandemic, type of vegetation, etc...)?  
+
+<|ref|>text<|/ref|><|det|>[[115, 203, 677, 218]]<|/det|>
+- > Thank you. We now try to make it clear – and the narrative has been toned down.  
+
+<|ref|>text<|/ref|><|det|>[[115, 247, 237, 262]]<|/det|>
+Interesting paper!  
+
+<|ref|>text<|/ref|><|det|>[[115, 300, 208, 315]]<|/det|>
+- > Thank you!  
+
+<|ref|>sub_title<|/ref|><|det|>[[120, 365, 282, 380]]<|/det|>
+## Final Decision Letter:  
+
+<|ref|>text<|/ref|><|det|>[[115, 399, 213, 414]]<|/det|>
+Dear Alessia,  
+
+<|ref|>text<|/ref|><|det|>[[115, 428, 872, 473]]<|/det|>
+We are pleased to inform you that your Article entitled "Palaeoecological data indicates land- use changes across Europe linked to spatial heterogeneity in mortality during the Black Death pandemic", has now been accepted for publication in Nature Ecology & Evolution.  
+
+<|ref|>text<|/ref|><|det|>[[115, 488, 874, 548]]<|/det|>
+Over the next few weeks, your paper will be copyedited to ensure that it conforms to Nature Ecology and Evolution style. Once your paper is typeset, you will receive an email with a link to choose the appropriate publishing options for your paper and our Author Services team will be in touch regarding any additional information that may be required  
+
+<|ref|>text<|/ref|><|det|>[[115, 563, 865, 608]]<|/det|>
+After the grant of rights is completed, you will receive a link to your electronic proof via email with a request to make any corrections within 48 hours. If, when you receive your proof, you cannot meet this deadline, please inform us at rjsproduction@springernature.com immediately.  
+
+<|ref|>text<|/ref|><|det|>[[115, 622, 872, 638]]<|/det|>
+You will not receive your proofs until the publishing agreement has been received through our system  
+
+<|ref|>text<|/ref|><|det|>[[115, 652, 880, 728]]<|/det|>
+Due to the importance of these deadlines, we ask you please us know now whether you will be difficult to contact over the next month. If this is the case, we ask you provide us with the contact information (email, phone and fax) of someone who will be able to check the proofs on your behalf, and who will be available to address any last- minute problems. Once your paper has been scheduled for online publication, the Nature press office will be in touch to confirm the details.  
+
+<|ref|>text<|/ref|><|det|>[[115, 742, 857, 802]]<|/det|>
+Acceptance of your manuscript is conditional on all authors' agreement with our publication policies (see www.nature.com/authors/policies/index.html). In particular your manuscript must not be published elsewhere and there must be no announcement of the work to any media outlet until the publication date (the day on which it is uploaded onto our web site).  
+
+<|ref|>text<|/ref|><|det|>[[115, 816, 875, 861]]<|/det|>
+Please note that <i>Nature Ecology & Evolution</i> is a Transformative Journal (TJ). Authors may publish their research with us through the traditional subscription access route or make their paper immediately open access through payment of an article- processing charge (APC). Authors will not be
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 50, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[115, 142, 860, 188]]<|/det|>
+required to make a final decision about access to their article until it has been accepted. <a href="https://www.springernature.com/gp/open- research/transformative- journals"> Find out more about Transformative Journals</a>  
+
+<|ref|>text<|/ref|><|det|>[[115, 203, 878, 368]]<|/det|>
+<B>Authors may need to take specific actions to achieve <a href="https://www.springernature.com/gp/open- research/funding/policy- compliance- faqs"> compliance</a> with funder and institutional open access mandates.</b> For submissions from January 2021, if your research is supported by a funder that requires immediate open access (e.g. according to <a href="https://www.springernature.com/gp/open- research/plan- s- compliance">Plan S principles</a>) then you should select the gold OA route, and we will direct you to the compliant route where possible. For authors selecting the subscription publication route our standard licensing terms will need to be accepted, including our <a href="https://www.springernature.com/gp/open- research/policies/journal- policies">self- archiving policies</a>. Those standard licensing terms will supersede any other terms that the author or any third party may assert apply to any version of the manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[115, 381, 843, 427]]<|/det|>
+In approximately 10 business days you will receive an email with a link to choose the appropriate publishing options for your paper and our Author Services team will be in touch regarding any additional information that may be required.  
+
+<|ref|>text<|/ref|><|det|>[[115, 441, 875, 457]]<|/det|>
+You will not receive your proofs until the publishing agreement has been received through our system.  
+
+<|ref|>text<|/ref|><|det|>[[115, 470, 875, 501]]<|/det|>
+If you have any questions about our publishing options, costs, Open Access requirements, or our legal forms, please contact ASJournals@springernature.com  
+
+<|ref|>text<|/ref|><|det|>[[115, 515, 596, 530]]<|/det|>
+An online order form for reprints of your paper is available at <a  
+
+<|ref|>text<|/ref|><|det|>[[115, 531, 870, 592]]<|/det|>
+href="https://www.nature.com/reprints/author- reprints.html">https://www.nature.com/reprints/author- reprints.html</a>. All co- authors, authors' institutions and authors' funding agencies can order reprints using the form appropriate to their geographical region.  
+
+<|ref|>text<|/ref|><|det|>[[115, 604, 880, 725]]<|/det|>
+We welcome the submission of potential cover material (including a short caption of around 40 words) related to your manuscript; suggestions should be sent to Nature Ecology & Evolution as electronic files (the image should be 300 dpi at \(210 \times 297 \text{mm}\) in either TIFF or JPEG format). Please note that such pictures should be selected more for their aesthetic appeal than for their scientific content, and that colour images work better than black and white or grayscale images. Please do not try to design a cover with the Nature Ecology & Evolution logo etc., and please do not submit composites of images related to your work. I am sure you will understand that we cannot make any promise as to whether any of your suggestions might be selected for the cover of the journal.  
+
+<|ref|>text<|/ref|><|det|>[[115, 739, 872, 784]]<|/det|>
+You can now use a single sign- on for all your accounts, view the status of all your manuscript submissions and reviews, access usage statistics for your published articles and download a record of your refereeing activity for the Nature journals.  
+
+<|ref|>text<|/ref|><|det|>[[115, 798, 878, 858]]<|/det|>
+To assist our authors in disseminating their research to the broader community, our SharedIt initiative provides you with a unique shareable link that will allow anyone (with or without a subscription) to read the published article. Recipients of the link with a subscription will also be able to download and print the PDF.
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[548, 48, 872, 85]]<|/det|>
+# natureresearch  
+
+<|ref|>text<|/ref|><|det|>[[115, 157, 833, 188]]<|/det|>
+You can generate the link yourself when you receive your article DOI by entering it here: <a href="http://authors.springernature.com/share">http://authors.springernature.com/share<a>.  
+
+<|ref|>text<|/ref|><|det|>[[115, 204, 234, 219]]<|/det|>
+Yours sincerely,  
+
+<|ref|>text<|/ref|><|det|>[[115, 248, 216, 263]]<|/det|>
+{REDACTED}
+
+<--- Page Split --->

peer_reviews/122b72a9723a3c9132ec90ab7bdc0cbef880f2e545ab1ae7ee8a6ffcabe67710/supplementary_0_Transparent Peer Review file/images_list.json

@@ -0,0 +1 @@
+[]

peer_reviews/122b72a9723a3c9132ec90ab7bdc0cbef880f2e545ab1ae7ee8a6ffcabe67710/supplementary_0_Transparent Peer Review file/supplementary_0_Transparent Peer Review file.mmd

@@ -0,0 +1,221 @@
+
+# Sex difference in BAT thermogenesis depends on PGC-1α- mediated phospholipid synthesis in mice  
+
+Corresponding Author: Dr Kazutaka Tsujimoto  
+
+This file contains all reviewer reports in order by version, followed by all author rebuttals in order by version.  
+
+Version 0:  
+
+Reviewer comments:  
+
+Reviewer #1  
+
+(Remarks to the Author)  
+
+In this manuscript, the authors determined the role of transcriptional coactivator PGC- 1α in sex- specific regulation of brown adipose tissue (BAT) thermogenic activity by investigating adipocyte- specific PGC- 1α knockout (AKO) male and female mice. They found that PGC- 1α AKO females exhibit less cold tolerance along with reduced BAT temperature and lower response to norepinephrine compared to males. This reduction in BAT thermogenesis was associated with altered mitochondrial membrane structure and lipid profiles, and decreased TCA cycle metabolites. In addition, expression of genes involved in ChREBPβ- mediated de novo lipogenesis (DNL) pathway was reduced in PGC- 1α AKO females but not in PGC- 1α AKO males. Based on these findings, the authors conclude that PGC- 1α plays a unique, sex- specific role in BAT of female mice through activation of ChREBPβ- mediated DNL, which contributes to the maintenance of mitochondrial membrane structure. However, it is not clear how PGC- 1α promotes ChREBPβ- mediated DNL. In addition, although the authors concluded that estrogen signaling also regulates ChREBPβ/DNL- related gene expression in female BAT, the relative contribution of direct estrogen signaling vs increased sympathetic nervous system (SNS) output to enhancing this pathway in BAT remains unclear. It is likely that PGC- 1α is one of many downstream factors activated in BAT of female mice due to BAT activation via enhanced SNS output.  
+
+1. It is not clear how PGC-1α promotes ChREBPβ-mediated DNL. Does PGC-1α interact with and coactivate ChREBPβ? PGC-1β has been shown to coactivate ChREBPβ and promote DNL gene expression. Does PGC-1α deletion alter PGC-1β expression?  
+
+2. It is not clear whether AAV-shChrebp targets specifically Chrebpb without altering Chrebpa expression. What is the effect of ChREBPβ knockdown on mitochondrial membrane structure and lipid profiles in male BAT?  
+
+3. Estrogen has been shown to enhance BAT function by increasing sympathetic nervous system (SNS) output to BAT (PMC4082097). In addition, mild cold (room temp)-induced AKT2 signaling has been shown to stimulate ChREBP-mediated DNL (PMC5762420). These findings indicate that increased SNS output to BAT in female mice can stimulate PGC-1α expression and ChRRBP-mediated DNL. Estrogen receptor antagonist TMX is likely to suppress SNS output to BAT. Thus, it is not clear whether TMX-mediated effect on BAT function is directly due to impaired estrogen signaling in BAT or due to reduced SNS output to BAT. Immunolabeling of tyrosine hydroxylase, a marker for sympathetic neurons, in male and female BAT may be considered.  
+
+## Reviewer #2  
+
+(Remarks to the Author)  
+
+In this study, the authors investigate the role mechanism(s) of PGC- 1α in sexual dimorphism. Inducible adipocyte- specific PGC- 1α knockout (KO) mice displayed decreased BAT thermogenesis only in females. Expression of Chrebpβ and downstream de novo lipogenesis (DNL) related genes were both reduced only in female KO mice. BAT- specific knockdown of Chrebpβ reduced the DNL- related gene expression and BAT thermogenesis in female wild- type mice. Furthermore, PGC- 1α enhanced the sensitivity of female BAT estrogen signaling, thereby increasing Chrebpβ and its downstream DNL- related gene expression. These findings suggest that PGC- 1α- ChREBPβ mediated DNL plays a pivotal role in BAT thermogenesis in a sex- dependent manner.  
+
+Major points  
+
+1- The stage of the estrus cycle is critical for Interpretation of the date obtained in female mice.
+
+<--- Page Split --->
+
+2- Is UCP1 expression more elevated in BAT from females than from males ?3- In Figure 3 and in Figure 7, the expression of ChREBPa (RT- qPCR) as well as ChREBP protein content should be measured (actually in males and in females)4- What are the functional consequences of the OVX- treatment on mitochondrial shape, /structure/function ?5- FGF21, a key actor of BAT function, known to be regulated in a sexual dimorphic manner should be measured in all the experiments presented.  
+
+## Reviewer #3  
+
+(Remarks to the Author)  
+
+The manuscript by Takeuchi et al found a female- specific effect of BAT thermogenesis most likely depending on PGC- 1a- CHREBPb mediated de novo lipogenesis. Overall, this study is well designed and executed and the findings are quite interesting. The major concern for this study is that lack of functional study on the DNL with only gene expressions of some DNL related genes, such as Acss2, FASN etc. It remains unknown why DNL in BAT is important for thermogenesis. Although the authors attempted to show the lipidomic profiles in BAT, especially cardiolipins. However, tetra linoleoyl CL is the major functional cardiolipins, and it is NOT dependent on DNL, since linoleic acid is an essential acid and can only be taken up from dietary sources. It is recommended to show some data that DNL is in fact important in the proposed mechanism using isotope tracing experiments, such as 3H labeled glucose, or 3H H2O. Below are some minor concerns:  
+
+1. Please indicate the weights/mass of BAT in different mice.  
+
+2. Figure 1d and 1f, Male KO mice also showed similar trends with female mice although it is not statistically significant. Is it because of the larger individual variation?  
+
+3. How to explain the difference of GTT in male and female mice in Fig.s1f  
+
+4. IN Figure 4A, please justify why they only focused on polar metabolites using CE-MS methods. IN the heatmap, how do they normalize the different metabolites?  
+
+5. In Figure 5e and 5f, why the patterns of CL(18:2) are different in these two figures?  
+
+6. In line 644, what does "10 uL brown adipose tissue samples" mean?  
+
+Version 1:  
+
+Reviewer comments:  
+
+Reviewer #1  
+
+(Remarks to the Author)  
+
+Reviewer #2  
+
+(Remarks to the Author)  
+
+No more comments  
+
+My queries were addressed, congratulations on the mice work !  
+
+Reviewer #3  
+
+(Remarks to the Author)  
+
+The authors satisfactorily addressed my concerns
+
+<--- Page Split --->
+
+Open Access This Peer Review File is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.  
+
+In cases where reviewers are anonymous, credit should be given to 'Anonymous Referee' and the source.  
+
+The images or other third party material in this Peer Review File are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.  
+
+To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/
+
+<--- Page Split --->
+
+Reviewer #1  
+
+## It is not clear how PGC-1α promotes ChREBPβ-mediated DNL. Does PGC-1α interact with and coactivate ChREBPβ?  
+
+It is not clear how PGC- 1α promotes ChREBPβ- mediated DNL. Does PGC- 1α interact with and coactivate ChREBPβ?We appreciate the insightful comments provided by Reviewer #1 and the editor, which have highlighted the critical importance of elucidating the role of PGC- 1α in ChREBPβ- mediated DNL. We fully recognize the significance of addressing this question and have conducted additional analyses to characterize the potential interaction between PGC- 1α and ChREBPβ. We showed that PGC- 1α KO reduced Chrebβ expression at the transcriptional level in female BAT. Although the direct regulation of Chrebβ transcription by PGC- 1α has not been reported, PGC- 1α is known to influence the transcription of target genes through various histone modifications (PMID: 19008463). Therefore, we performed ATAC- seq analysis to assess chromatin accessibility at the relevant regulatory regions. Our results showed that the female BAT exhibited greater chromatin accessibility near the transcription start site (TSS) of Chrebβ than the male BAT (Fig. 3g), suggesting a sex- specific regulatory mechanism. Notably, this sex difference in chromatin accessibility was abolished only in female PGC- 1α knockout (KO) mice. These results indicate that PGC- 1α in female BAT regulates ChREBPβ expression at the transcriptional level through the modulation of chromatin accessibility in a female- specific manner, rather than through direct interaction. Although further examinations are required to clarify the mechanism by which PGC- 1α modulates chromatin accessibility near the TSS of Chrebβ only in female mice, we will be committed to addressing this important question in our future studies. These results have been added to the Results section (Lines 171–189, Figure 3d–g).  
+
+These results have been added to the Results section (Lines 171- 189, Figure 3d- g).  
+
+## PGC-1β has been shown to coactivate ChREBPβ and promote DNL gene expression.  
+
+## Does PGC-1α deletion alter PGC-1β expression?  
+
+To answer this question, we additionally examined Pgc1β expression levels in BAT. No significant difference in Pgc1β expression was observed between male and female BAT. Moreover, Pgc1β expression levels were not affected by PGC- 1α knockout (Fig. S1g). These results indicate that PGC- 1α regulates expression of ChREBPβ and DNL- related genes in a PGC- 1β- independent manner.  
+
+These results have been added to the Results section of the revised manuscript (Lines 104- 105, Supplementary Figure 1g).  
+
+It is not clear whether AAV- shChrebp targets specifically Chrebpb without altering Chrebpa expression. What is the effect of ChREBPβ knockdown on mitochondrial membrane structure and lipid profiles in male BAT?
+
+<--- Page Split --->
+
+We evaluated the knockdown (KD) efficiency of AAV- shChrebp on Chrebpβ and Chrebpa expression in the BAT of both sexes. Our results demonstrated that AAV- shChrebp administration almost completely suppressed Chrebpβ expression while having no significant effect on Chrebpa expression in both male and female BAT (Fig. 6a, Fig. S5a, Fig. S6a- b). These results indicate that AAV- shChrebp selectively targets Chrebpβ expression in the BAT of male and female mice.  
+
+These results have been added to the Results section of the revised manuscript (Lines 252- 254: Figure 6a; Supplementary Figure 5a, and Lines 273- 274: Supplementary Figures 6a- b).  
+
+The reviewer requested examination of the effects of ChREBPβ KD on mitochondrial membrane structure and lipid profiles in male BAT. Furthermore, it appears that investigating the effects of ChREBPβ KD on lipid profiles in female BAT is necessary to assess its effect on the mitochondrial membrane structure. Therefore, we additionally analyzed the lipid profile in the BAT of female ChREBPβ KD mice.  
+
+We first examined the effect of ChREBPβ KD on mitochondrial membrane structure and lipid profiles in male BAT. In male mice, ChREBPβ KD did not affect the mitochondrial membrane structure (Fig. S6c- e). Therefore, we focused on the levels of cardiolipin (CL) and ether- linked phosphatidylethanolamines (PEs), which are known to maintain mitochondrial morphology and function (PMID: 29034233, 37069167, 38129691). Notably, CL(18:2)4 and ether- linked PEs were found to be reduced only in female Pgc- 1α KO mice (Fig. 5f- g, Fig. S5e). Interestingly, these lipid levels were not reduced in the BAT of male ChREBPβ KD mice (Fig. S6f- g).  
+
+Next, we examined the effects of ChREBPβ KD on CL(18:2)4 and ether- linked PE levels in female BAT. In female BAT, ChREBPβ KD reduced ether- linked PE levels but did not affect CL(18:2)4 levels (Fig. 6f- g). These results indicate that a female- specific ChREBPβ- dependent mechanism may enhance ether- linked PE levels in BAT, which could play a critical role in maintaining mitochondrial morphology and function. In contrast, ChREBPβ did not appear to be involved in PGC- 1α- mediated CL(18:2)4 production in female BAT.  
+
+In addition, we observed that PGC- 1α KO did not affect Chrebpa expression in the BAT of either sex (Fig. S3a). Overall, these findings indicate that PGC- 1α may regulate mitochondrial membrane structure through ChREBPβ- dependent ether- linked PE production and ChREBPβ- independent CL(18:2)4 production in a female- specific manner. These results have been added to the Results section of the revised manuscript (Lines 273- 279: Supplementary Figure 6a- g, and Lines 263- 266: Figure 6f- g) and discussed in the Discussion section of the revised manuscript (Lines 370- 373 and Lines 379- 386).
+
+<--- Page Split --->
+
+Estrogen has been shown to enhance BAT function by increasing sympathetic nervous system (SNS) output to BAT (PMC4082097). In addition, mild cold (room temp)- induced AKT2 signaling has been shown to stimulate ChREBP- mediated DNL (PMC5762420). These findings indicate that increased SNS output to BAT in female mice can stimulate PGC- 1α expression and ChREBP- mediated DNL. Estrogen receptor antagonist TMX is likely to suppress SNS output to BAT. Thus, it is not clear whether TMX- mediated effect on BAT function is directly due to impaired estrogen signaling in BAT or due to reduced SNS output to BAT. Immunolabeling of tyrosine hydroxylase, a marker for sympathetic neurons, in male and female BAT may be considered.  
+
+We appreciate the reviewer's insightful comment regarding the potential role of estrogen signaling and SNS output in regulating BAT function. To address this concern, we performed the immunolabeling of tyrosine hydroxylase (TH) and examined the mRNA levels of TH in the BAT of male and female wild- type mice treated with either TMX or vehicle for 5 days. Our results showed no significant differences in TH expression between the sexes or between the TMX and vehicle groups (Fig. S8a- b). These findings suggest that TMX- induced estrogen antagonism does not significantly affect SNS output to BAT in females. Furthermore, as shown in Figure 7g, TMX administration did not alter BAT Pgc1α gene expression, further supporting the conclusion that TMX does not influence SNS output. These results have been added to the Results section of the revised manuscript (Lines 307- 312: Supplementary Figure 8a- b).  
+
+## Reviewer #2  
+
+## The stage of the estrus cycle is critical for Interpretation of the date obtained in female mice.  
+
+As the reviewer pointed out, the estrus cycle might influence the interpretation of the experimental results. However, in our study, consistent phenotypes—including suppressed thermogenesis and mitochondrial structural deficits in female BAT, induced by either PGC- 1α KO or TMX treatment—were observed across all experiments conducted on a single experimental day (Fig. 1- 7), without any special consideration of the estrus cycle. To determine why this consistency was achieved, we collected blood samples from wild- type female mice during a single experimental day, as in our previous experiments, and measured serum estrogen levels. As shown in the figure below, we observed variability in serum estrogen levels among individual mice, which likely reflects the different stages of the estrus cycle on a single experimental day. Importantly, all experiments (Fig. 1- 7) were performed under these conditions, where female mice had varying serum estrogen levels.
+
+<--- Page Split --->
+
+Despite this variability, consistent phenotypes were obtained, suggesting that the observed effects are not dependent on the estrus cycle stage. Furthermore, serum estrogen levels in male mice were found to be below the detection limit, consistent with a previous report (PMID: 25856427). These findings collectively suggest that even the lowest levels of estrogen in female mice are sufficient to account for the observed sex differences in BAT thermogenesis. Therefore, the estrus cycle stage may not be critical for interpreting the results obtained from female mice.  
+
+![PLACEHOLDER_6_0]
+  
+
+## Is UCP1 expression more elevated in BAT from females than from males?  
+
+We measured the expression levels of Ucp1 in BAT. The expression levels of Ucp1 in BAT were higher in females than in males (Fig. S1a), consistent with the enhancement of BAT thermogenesis in female mice.  
+
+These results have been added to the Results section of the revised manuscript (Lines 88- 89: Supplementary Figure 1a).  
+
+## In Figure 3 and in Figure 7, the expression of ChREBPa (RT-qPCR) as well as ChREBP protein content should be measured (actually in males and in females)  
+
+We measured Chrebpa gene expression levels and ChREBP protein levels in the BAT of both PGC- 1α KO mice and TMX- , an estrogen receptor antagonist, administered mice. Neither PGC- 1α KO (Fig. S3a) nor TMX administration (Fig. S7b) resulted in any changes in the expression levels of Chrebpa in either sex. Moreover, the expression levels of the ChREBP protein in BAT were higher in females than in males. In females, both PGC- 1α KO (Fig. S3b) and TMX administration (Fig. S7c) reduced BAT ChREBP protein expression levels, whereas in males, neither PGC- 1α KO nor TMX treatment induced such reductions.  
+
+These results have been added to the Results section of the revised manuscript (Lines 162- 166: Supplementary Figure 3a- b, and Lines 294- 296: Supplementary Figure 7b- c).  
+
+What are the functional consequences of the OVX- treatment on mitochondrial shape,
+
+<--- Page Split --->
+
+## /structure/function ?  
+
+In the BAT of ovariectomized mice, the mitochondrial area was smaller, and the total length of the cristae was also reduced compared with that in the sham- operated female mice (Fig. S7g- i). In addition, VO2 after NE administration was reduced in ovariectomized mice compared with that in sham- operated female mice (Fig. S7f). These results were similar to those observed in the TMX- administered female mice in this study. Overall, these results suggest that estrogen signaling plays an important role in BAT thermogenesis in female mice.  
+
+These results have been added to the Results section of the revised manuscript (Lines 302- 306: Supplementary Figure 7f- i).  
+
+## FGF21, a key actor of BAT function, known to be regulated in a sexual dimorphic manner should be measured in all the experiments presented.  
+
+We measured the expression levels of Fgf21 in BAT in all experiments. No significant differences were observed in the expression levels of Fgf21 between the control and intervention groups as well as between males and females (Fig. S3c, and as shown in the figure below). These results suggest that FGF21 did not play an important role in the sex difference in BAT thermogenesis in this model.  
+
+These results have been added to the Results section of the revised manuscript (Lines 166- 168: Supplementary Figure 3c).  
+
+![PLACEHOLDER_7_0]
+  
+
+Reviewer #3  
+
+However, tetra linoleoyl CL is the major functional cardiolipins, and it is NOT dependent on DNL, since linoleic acid is an essential acid and can only be taken up from dietary sources. It is recommended to show some data that DNL is in fact important in the proposed mechanism using isotope tracing experiments, such as 3H labeled glucose, or 3H H2O.  
+
+We deeply appreciate your valuable advice and suggestions. As suggested, we performed
+
+<--- Page Split --->
+
+lipidomics after \(\mathsf{D}_2\mathsf{O}\) administration and found that the labeled components of CL(18:2)4 were not detected in the PGC- 1α KO and Control mice of both sexes.. This result clearly indicates that PGC- 1α does not enhance CL(18:2)4 production through DNL in the BAT of female mice. In addition, we found that ChREBPβ knockdown did not affect CL(18:2)4 levels in female BAT (Fig. 6f). Overall, these results suggest that PGC- 1α increases CL(18:2)4 levels through molecular mechanism(s) other than ChREBPβ- mediated DNL in female BAT. In addition, we found that several molecular species of ether- linked phosphatidylethanolamine (PE), which were decreased in the female BAT of PGC- 1α KO mice (Fig. S5e), were commonly reduced following BAT ChREBPβ knockdown in female mice (Fig. 6g). The labeled components of ether- linked PE(O- 16:1_18:1) showed a significant decrease in lipidomics after \(\mathsf{D}_2\mathsf{O}\) administration in female Pgc- 1α KO mice (Fig. 6h). These results suggest that ether- linked PE is synthesized via PGC- 1α- ChREBPβ- mediated DNL. Recent studies have highlighted that ether- linked phospholipids play a crucial role in maintaining the morphology and function of the mitochondria (PMID: 37069167, 38129691).  
+
+In summary, PGC- 1α may play a pivotal role in the mitochondrial function of female BAT through both ChREBPβ- dependent de novo ether- linked PE production and ChREBPβ- independent CL(18:2)4 production.  
+
+These observations have been added to the Results section of the revised manuscript (Lines 263- 272: Figure 6f- h, Supplementary Figure 5e) and discussed in the Discussion section of the revised manuscript (Lines 370- 373 and Lines 379- 386).  
+
+## Below are some minor concerns:  
+
+## 1. Please indicate the weights/mass of BAT in different mice.  
+
+We measured the BAT mass in PGC- 1α KO or TMX-, an estrogen receptor antagonist, administered mice. As shown in the figure below, neither PGC- 1α KO nor TMX administration resulted in any change in BAT weight in either sex.  
+
+![PLACEHOLDER_8_0]
+
+
+<--- Page Split --->
+
+## 2. Figure 1d and 1f, Male KO mice also showed similar trends with female mice although it is not statistically significant. Is it because of the larger individual variation?  
+
+We agree that body temperature during cold exposure tended to be lower in Male KO mice than in Male Control mice. However, as the reviewer pointed out, these differences were not statistically significant. Moreover, the effects of PGC- 1α KO on BAT, such as energy expenditure (Figure 1g), gene expression (Figure 3a- c), metabolite profile (Figure 4a- e), lipid profile (Figure 5a- i), and mitochondrial structure (Figure 2a- c) and activity (Figure 2d and Figure 4e), were all observed only in female BAT, which is completely consistent with the fact that cold tolerance was significantly reduced only in female mice (Figure 1d, f). Therefore, the lack of a significant reduction in cold tolerance in male mice may be due to sex differences in BAT function.  
+
+## 3. How to explain the difference of GTT in male and female mice in Fig.s1f  
+
+A previous study (PMID: 22645355) reported that male adipocyte- specific PGC- 1α KO mice were systemic glucose intolerant. Our results are compatible with those of this previous report. Given that male mice have poorer glucose tolerance than female mice, impaired glucose tolerance may be more pronounced in male mice than in female mice, although further studies are required to clarify the mechanisms underlying this sex difference.  
+
+## 4. IN Figure 4A, please justify why they only focused on polar metabolites using CEMS methods. IN the heatmap, how do they normalize the different metabolites?  
+
+We focused only on polar metabolites because most molecules produced when the mitochondrial TCA cycle and electron transport chain are activated during BAT thermogenesis are water- soluble and easily ionized molecules. CE- MS methods were used because they are suitable for analyzing polar metabolites. In the heatmap analysis, we used the following formula for normalization: \(Z = (X - \mu) / \sigma\) (Z: standardized value of each metabolite, X: Relative area of each metabolite, \(\mu\) : mean of each metabolite, \(\sigma\) : SD of each metabolite)  
+
+## 5. In Figure 5e and 5f, why the patterns of CL(18:2) are different in these two figures?  
+
+Figure 5e shows the proportion of each CL molecular species, including CL(18:2)4, relative to the total lipid content in BAT. In contrast, Figure 5f presents the proportion of the CL(18:2)4 molecular species normalized to the total CL content in BAT. The proportions of CL among
+
+<--- Page Split --->
+
+all lipids are very small, and those of individual CL species are greatly influenced not only by changes in the absolute amount of individual CL species but also by changes in the absolute amount of total lipids excluding CL. Therefore, to evaluate the changes in CL(18:2)4 within CL, we measured the proportion of CL(18:2)4 among CL as previously reported (PMID: 16855048) and found that CL(18:2)4 was actually reduced only in female knockout mice, as shown in Figure 5f.  
+
+## 6. In line 644, what does "10 uL brown adipose tissue samples" mean?  
+
+We thank the reviewer for this comment. This is a typographical error; the correct notation should be "1 mg brown adipose tissue." We have corrected this error in the text.
+
+<--- Page Split --->

peer_reviews/122b72a9723a3c9132ec90ab7bdc0cbef880f2e545ab1ae7ee8a6ffcabe67710/supplementary_0_Transparent Peer Review file/supplementary_0_Transparent Peer Review file_det.mmd

@@ -0,0 +1,302 @@
+<|ref|>title<|/ref|><|det|>[[73, 161, 891, 211]]<|/det|>
+# Sex difference in BAT thermogenesis depends on PGC-1α- mediated phospholipid synthesis in mice  
+
+<|ref|>text<|/ref|><|det|>[[73, 224, 471, 241]]<|/det|>
+Corresponding Author: Dr Kazutaka Tsujimoto  
+
+<|ref|>text<|/ref|><|det|>[[72, 275, 864, 290]]<|/det|>
+This file contains all reviewer reports in order by version, followed by all author rebuttals in order by version.  
+
+<|ref|>text<|/ref|><|det|>[[73, 327, 145, 340]]<|/det|>
+Version 0:  
+
+<|ref|>text<|/ref|><|det|>[[73, 354, 220, 367]]<|/det|>
+Reviewer comments:  
+
+<|ref|>text<|/ref|><|det|>[[73, 380, 160, 393]]<|/det|>
+Reviewer #1  
+
+<|ref|>text<|/ref|><|det|>[[73, 405, 238, 418]]<|/det|>
+(Remarks to the Author)  
+
+<|ref|>text<|/ref|><|det|>[[72, 418, 920, 589]]<|/det|>
+In this manuscript, the authors determined the role of transcriptional coactivator PGC- 1α in sex- specific regulation of brown adipose tissue (BAT) thermogenic activity by investigating adipocyte- specific PGC- 1α knockout (AKO) male and female mice. They found that PGC- 1α AKO females exhibit less cold tolerance along with reduced BAT temperature and lower response to norepinephrine compared to males. This reduction in BAT thermogenesis was associated with altered mitochondrial membrane structure and lipid profiles, and decreased TCA cycle metabolites. In addition, expression of genes involved in ChREBPβ- mediated de novo lipogenesis (DNL) pathway was reduced in PGC- 1α AKO females but not in PGC- 1α AKO males. Based on these findings, the authors conclude that PGC- 1α plays a unique, sex- specific role in BAT of female mice through activation of ChREBPβ- mediated DNL, which contributes to the maintenance of mitochondrial membrane structure. However, it is not clear how PGC- 1α promotes ChREBPβ- mediated DNL. In addition, although the authors concluded that estrogen signaling also regulates ChREBPβ/DNL- related gene expression in female BAT, the relative contribution of direct estrogen signaling vs increased sympathetic nervous system (SNS) output to enhancing this pathway in BAT remains unclear. It is likely that PGC- 1α is one of many downstream factors activated in BAT of female mice due to BAT activation via enhanced SNS output.  
+
+<|ref|>text<|/ref|><|det|>[[72, 600, 920, 640]]<|/det|>
+1. It is not clear how PGC-1α promotes ChREBPβ-mediated DNL. Does PGC-1α interact with and coactivate ChREBPβ? PGC-1β has been shown to coactivate ChREBPβ and promote DNL gene expression. Does PGC-1α deletion alter PGC-1β expression?  
+
+<|ref|>text<|/ref|><|det|>[[72, 640, 920, 667]]<|/det|>
+2. It is not clear whether AAV-shChrebp targets specifically Chrebpb without altering Chrebpa expression. What is the effect of ChREBPβ knockdown on mitochondrial membrane structure and lipid profiles in male BAT?  
+
+<|ref|>text<|/ref|><|det|>[[72, 667, 923, 758]]<|/det|>
+3. Estrogen has been shown to enhance BAT function by increasing sympathetic nervous system (SNS) output to BAT (PMC4082097). In addition, mild cold (room temp)-induced AKT2 signaling has been shown to stimulate ChREBP-mediated DNL (PMC5762420). These findings indicate that increased SNS output to BAT in female mice can stimulate PGC-1α expression and ChRRBP-mediated DNL. Estrogen receptor antagonist TMX is likely to suppress SNS output to BAT. Thus, it is not clear whether TMX-mediated effect on BAT function is directly due to impaired estrogen signaling in BAT or due to reduced SNS output to BAT. Immunolabeling of tyrosine hydroxylase, a marker for sympathetic neurons, in male and female BAT may be considered.  
+
+<|ref|>sub_title<|/ref|><|det|>[[73, 783, 161, 795]]<|/det|>
+## Reviewer #2  
+
+<|ref|>text<|/ref|><|det|>[[73, 810, 238, 822]]<|/det|>
+(Remarks to the Author)  
+
+<|ref|>text<|/ref|><|det|>[[72, 823, 920, 913]]<|/det|>
+In this study, the authors investigate the role mechanism(s) of PGC- 1α in sexual dimorphism. Inducible adipocyte- specific PGC- 1α knockout (KO) mice displayed decreased BAT thermogenesis only in females. Expression of Chrebpβ and downstream de novo lipogenesis (DNL) related genes were both reduced only in female KO mice. BAT- specific knockdown of Chrebpβ reduced the DNL- related gene expression and BAT thermogenesis in female wild- type mice. Furthermore, PGC- 1α enhanced the sensitivity of female BAT estrogen signaling, thereby increasing Chrebpβ and its downstream DNL- related gene expression. These findings suggest that PGC- 1α- ChREBPβ mediated DNL plays a pivotal role in BAT thermogenesis in a sex- dependent manner.  
+
+<|ref|>text<|/ref|><|det|>[[73, 914, 160, 925]]<|/det|>
+Major points  
+
+<|ref|>text<|/ref|><|det|>[[72, 926, 713, 939]]<|/det|>
+1- The stage of the estrus cycle is critical for Interpretation of the date obtained in female mice.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[72, 46, 901, 125]]<|/det|>
+2- Is UCP1 expression more elevated in BAT from females than from males ?3- In Figure 3 and in Figure 7, the expression of ChREBPa (RT- qPCR) as well as ChREBP protein content should be measured (actually in males and in females)4- What are the functional consequences of the OVX- treatment on mitochondrial shape, /structure/function ?5- FGF21, a key actor of BAT function, known to be regulated in a sexual dimorphic manner should be measured in all the experiments presented.  
+
+<|ref|>sub_title<|/ref|><|det|>[[73, 164, 162, 177]]<|/det|>
+## Reviewer #3  
+
+<|ref|>text<|/ref|><|det|>[[73, 189, 238, 202]]<|/det|>
+(Remarks to the Author)  
+
+<|ref|>text<|/ref|><|det|>[[72, 202, 912, 288]]<|/det|>
+The manuscript by Takeuchi et al found a female- specific effect of BAT thermogenesis most likely depending on PGC- 1a- CHREBPb mediated de novo lipogenesis. Overall, this study is well designed and executed and the findings are quite interesting. The major concern for this study is that lack of functional study on the DNL with only gene expressions of some DNL related genes, such as Acss2, FASN etc. It remains unknown why DNL in BAT is important for thermogenesis. Although the authors attempted to show the lipidomic profiles in BAT, especially cardiolipins. However, tetra linoleoyl CL is the major functional cardiolipins, and it is NOT dependent on DNL, since linoleic acid is an essential acid and can only be taken up from dietary sources. It is recommended to show some data that DNL is in fact important in the proposed mechanism using isotope tracing experiments, such as 3H labeled glucose, or 3H H2O. Below are some minor concerns:  
+
+<|ref|>text<|/ref|><|det|>[[72, 288, 910, 312]]<|/det|>
+1. Please indicate the weights/mass of BAT in different mice.  
+
+<|ref|>text<|/ref|><|det|>[[72, 312, 912, 339]]<|/det|>
+2. Figure 1d and 1f, Male KO mice also showed similar trends with female mice although it is not statistically significant. Is it because of the larger individual variation?  
+
+<|ref|>text<|/ref|><|det|>[[72, 339, 581, 353]]<|/det|>
+3. How to explain the difference of GTT in male and female mice in Fig.s1f  
+
+<|ref|>text<|/ref|><|det|>[[72, 353, 904, 380]]<|/det|>
+4. IN Figure 4A, please justify why they only focused on polar metabolites using CE-MS methods. IN the heatmap, how do they normalize the different metabolites?  
+
+<|ref|>text<|/ref|><|det|>[[72, 380, 652, 395]]<|/det|>
+5. In Figure 5e and 5f, why the patterns of CL(18:2) are different in these two figures?  
+
+<|ref|>text<|/ref|><|det|>[[72, 395, 559, 409]]<|/det|>
+6. In line 644, what does "10 uL brown adipose tissue samples" mean?  
+
+<|ref|>text<|/ref|><|det|>[[73, 433, 144, 447]]<|/det|>
+Version 1:  
+
+<|ref|>text<|/ref|><|det|>[[73, 460, 220, 474]]<|/det|>
+Reviewer comments:  
+
+<|ref|>text<|/ref|><|det|>[[73, 486, 161, 500]]<|/det|>
+Reviewer #1  
+
+<|ref|>text<|/ref|><|det|>[[73, 512, 238, 527]]<|/det|>
+(Remarks to the Author)  
+
+<|ref|>text<|/ref|><|det|>[[73, 551, 162, 565]]<|/det|>
+Reviewer #2  
+
+<|ref|>text<|/ref|><|det|>[[73, 578, 238, 591]]<|/det|>
+(Remarks to the Author)  
+
+<|ref|>text<|/ref|><|det|>[[73, 592, 199, 604]]<|/det|>
+No more comments  
+
+<|ref|>text<|/ref|><|det|>[[73, 604, 506, 618]]<|/det|>
+My queries were addressed, congratulations on the mice work !  
+
+<|ref|>text<|/ref|><|det|>[[73, 631, 162, 644]]<|/det|>
+Reviewer #3  
+
+<|ref|>text<|/ref|><|det|>[[73, 657, 238, 670]]<|/det|>
+(Remarks to the Author)  
+
+<|ref|>text<|/ref|><|det|>[[73, 671, 415, 685]]<|/det|>
+The authors satisfactorily addressed my concerns
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[72, 45, 916, 99]]<|/det|>
+Open Access This Peer Review File is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.  
+
+<|ref|>text<|/ref|><|det|>[[72, 99, 796, 113]]<|/det|>
+In cases where reviewers are anonymous, credit should be given to 'Anonymous Referee' and the source.  
+
+<|ref|>text<|/ref|><|det|>[[72, 112, 910, 166]]<|/det|>
+The images or other third party material in this Peer Review File are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.  
+
+<|ref|>text<|/ref|><|det|>[[72, 166, 618, 180]]<|/det|>
+To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[140, 120, 236, 135]]<|/det|>
+Reviewer #1  
+
+<|ref|>sub_title<|/ref|><|det|>[[140, 140, 839, 179]]<|/det|>
+## It is not clear how PGC-1α promotes ChREBPβ-mediated DNL. Does PGC-1α interact with and coactivate ChREBPβ?  
+
+<|ref|>text<|/ref|><|det|>[[139, 183, 855, 580]]<|/det|>
+It is not clear how PGC- 1α promotes ChREBPβ- mediated DNL. Does PGC- 1α interact with and coactivate ChREBPβ?We appreciate the insightful comments provided by Reviewer #1 and the editor, which have highlighted the critical importance of elucidating the role of PGC- 1α in ChREBPβ- mediated DNL. We fully recognize the significance of addressing this question and have conducted additional analyses to characterize the potential interaction between PGC- 1α and ChREBPβ. We showed that PGC- 1α KO reduced Chrebβ expression at the transcriptional level in female BAT. Although the direct regulation of Chrebβ transcription by PGC- 1α has not been reported, PGC- 1α is known to influence the transcription of target genes through various histone modifications (PMID: 19008463). Therefore, we performed ATAC- seq analysis to assess chromatin accessibility at the relevant regulatory regions. Our results showed that the female BAT exhibited greater chromatin accessibility near the transcription start site (TSS) of Chrebβ than the male BAT (Fig. 3g), suggesting a sex- specific regulatory mechanism. Notably, this sex difference in chromatin accessibility was abolished only in female PGC- 1α knockout (KO) mice. These results indicate that PGC- 1α in female BAT regulates ChREBPβ expression at the transcriptional level through the modulation of chromatin accessibility in a female- specific manner, rather than through direct interaction. Although further examinations are required to clarify the mechanism by which PGC- 1α modulates chromatin accessibility near the TSS of Chrebβ only in female mice, we will be committed to addressing this important question in our future studies. These results have been added to the Results section (Lines 171–189, Figure 3d–g).  
+
+<|ref|>text<|/ref|><|det|>[[140, 576, 784, 594]]<|/det|>
+These results have been added to the Results section (Lines 171- 189, Figure 3d- g).  
+
+<|ref|>sub_title<|/ref|><|det|>[[140, 614, 833, 632]]<|/det|>
+## PGC-1β has been shown to coactivate ChREBPβ and promote DNL gene expression.  
+
+<|ref|>sub_title<|/ref|><|det|>[[140, 635, 537, 651]]<|/det|>
+## Does PGC-1α deletion alter PGC-1β expression?  
+
+<|ref|>text<|/ref|><|det|>[[140, 655, 850, 757]]<|/det|>
+To answer this question, we additionally examined Pgc1β expression levels in BAT. No significant difference in Pgc1β expression was observed between male and female BAT. Moreover, Pgc1β expression levels were not affected by PGC- 1α knockout (Fig. S1g). These results indicate that PGC- 1α regulates expression of ChREBPβ and DNL- related genes in a PGC- 1β- independent manner.  
+
+<|ref|>text<|/ref|><|det|>[[140, 760, 848, 800]]<|/det|>
+These results have been added to the Results section of the revised manuscript (Lines 104- 105, Supplementary Figure 1g).  
+
+<|ref|>text<|/ref|><|det|>[[140, 823, 825, 884]]<|/det|>
+It is not clear whether AAV- shChrebp targets specifically Chrebpb without altering Chrebpa expression. What is the effect of ChREBPβ knockdown on mitochondrial membrane structure and lipid profiles in male BAT?
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[139, 118, 850, 243]]<|/det|>
+We evaluated the knockdown (KD) efficiency of AAV- shChrebp on Chrebpβ and Chrebpa expression in the BAT of both sexes. Our results demonstrated that AAV- shChrebp administration almost completely suppressed Chrebpβ expression while having no significant effect on Chrebpa expression in both male and female BAT (Fig. 6a, Fig. S5a, Fig. S6a- b). These results indicate that AAV- shChrebp selectively targets Chrebpβ expression in the BAT of male and female mice.  
+
+<|ref|>text<|/ref|><|det|>[[140, 247, 848, 287]]<|/det|>
+These results have been added to the Results section of the revised manuscript (Lines 252- 254: Figure 6a; Supplementary Figure 5a, and Lines 273- 274: Supplementary Figures 6a- b).  
+
+<|ref|>text<|/ref|><|det|>[[139, 310, 850, 415]]<|/det|>
+The reviewer requested examination of the effects of ChREBPβ KD on mitochondrial membrane structure and lipid profiles in male BAT. Furthermore, it appears that investigating the effects of ChREBPβ KD on lipid profiles in female BAT is necessary to assess its effect on the mitochondrial membrane structure. Therefore, we additionally analyzed the lipid profile in the BAT of female ChREBPβ KD mice.  
+
+<|ref|>text<|/ref|><|det|>[[139, 418, 855, 586]]<|/det|>
+We first examined the effect of ChREBPβ KD on mitochondrial membrane structure and lipid profiles in male BAT. In male mice, ChREBPβ KD did not affect the mitochondrial membrane structure (Fig. S6c- e). Therefore, we focused on the levels of cardiolipin (CL) and ether- linked phosphatidylethanolamines (PEs), which are known to maintain mitochondrial morphology and function (PMID: 29034233, 37069167, 38129691). Notably, CL(18:2)4 and ether- linked PEs were found to be reduced only in female Pgc- 1α KO mice (Fig. 5f- g, Fig. S5e). Interestingly, these lipid levels were not reduced in the BAT of male ChREBPβ KD mice (Fig. S6f- g).  
+
+<|ref|>text<|/ref|><|det|>[[139, 589, 855, 712]]<|/det|>
+Next, we examined the effects of ChREBPβ KD on CL(18:2)4 and ether- linked PE levels in female BAT. In female BAT, ChREBPβ KD reduced ether- linked PE levels but did not affect CL(18:2)4 levels (Fig. 6f- g). These results indicate that a female- specific ChREBPβ- dependent mechanism may enhance ether- linked PE levels in BAT, which could play a critical role in maintaining mitochondrial morphology and function. In contrast, ChREBPβ did not appear to be involved in PGC- 1α- mediated CL(18:2)4 production in female BAT.  
+
+<|ref|>text<|/ref|><|det|>[[139, 716, 850, 862]]<|/det|>
+In addition, we observed that PGC- 1α KO did not affect Chrebpa expression in the BAT of either sex (Fig. S3a). Overall, these findings indicate that PGC- 1α may regulate mitochondrial membrane structure through ChREBPβ- dependent ether- linked PE production and ChREBPβ- independent CL(18:2)4 production in a female- specific manner. These results have been added to the Results section of the revised manuscript (Lines 273- 279: Supplementary Figure 6a- g, and Lines 263- 266: Figure 6f- g) and discussed in the Discussion section of the revised manuscript (Lines 370- 373 and Lines 379- 386).
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[137, 118, 856, 303]]<|/det|>
+Estrogen has been shown to enhance BAT function by increasing sympathetic nervous system (SNS) output to BAT (PMC4082097). In addition, mild cold (room temp)- induced AKT2 signaling has been shown to stimulate ChREBP- mediated DNL (PMC5762420). These findings indicate that increased SNS output to BAT in female mice can stimulate PGC- 1α expression and ChREBP- mediated DNL. Estrogen receptor antagonist TMX is likely to suppress SNS output to BAT. Thus, it is not clear whether TMX- mediated effect on BAT function is directly due to impaired estrogen signaling in BAT or due to reduced SNS output to BAT. Immunolabeling of tyrosine hydroxylase, a marker for sympathetic neurons, in male and female BAT may be considered.  
+
+<|ref|>text<|/ref|><|det|>[[138, 305, 857, 567]]<|/det|>
+We appreciate the reviewer's insightful comment regarding the potential role of estrogen signaling and SNS output in regulating BAT function. To address this concern, we performed the immunolabeling of tyrosine hydroxylase (TH) and examined the mRNA levels of TH in the BAT of male and female wild- type mice treated with either TMX or vehicle for 5 days. Our results showed no significant differences in TH expression between the sexes or between the TMX and vehicle groups (Fig. S8a- b). These findings suggest that TMX- induced estrogen antagonism does not significantly affect SNS output to BAT in females. Furthermore, as shown in Figure 7g, TMX administration did not alter BAT Pgc1α gene expression, further supporting the conclusion that TMX does not influence SNS output. These results have been added to the Results section of the revised manuscript (Lines 307- 312: Supplementary Figure 8a- b).  
+
+<|ref|>sub_title<|/ref|><|det|>[[140, 592, 238, 606]]<|/det|>
+## Reviewer #2  
+
+<|ref|>sub_title<|/ref|><|det|>[[140, 611, 802, 649]]<|/det|>
+## The stage of the estrus cycle is critical for Interpretation of the date obtained in female mice.  
+
+<|ref|>text<|/ref|><|det|>[[139, 653, 848, 884]]<|/det|>
+As the reviewer pointed out, the estrus cycle might influence the interpretation of the experimental results. However, in our study, consistent phenotypes—including suppressed thermogenesis and mitochondrial structural deficits in female BAT, induced by either PGC- 1α KO or TMX treatment—were observed across all experiments conducted on a single experimental day (Fig. 1- 7), without any special consideration of the estrus cycle. To determine why this consistency was achieved, we collected blood samples from wild- type female mice during a single experimental day, as in our previous experiments, and measured serum estrogen levels. As shown in the figure below, we observed variability in serum estrogen levels among individual mice, which likely reflects the different stages of the estrus cycle on a single experimental day. Importantly, all experiments (Fig. 1- 7) were performed under these conditions, where female mice had varying serum estrogen levels.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[139, 119, 852, 265]]<|/det|>
+Despite this variability, consistent phenotypes were obtained, suggesting that the observed effects are not dependent on the estrus cycle stage. Furthermore, serum estrogen levels in male mice were found to be below the detection limit, consistent with a previous report (PMID: 25856427). These findings collectively suggest that even the lowest levels of estrogen in female mice are sufficient to account for the observed sex differences in BAT thermogenesis. Therefore, the estrus cycle stage may not be critical for interpreting the results obtained from female mice.  
+
+<|ref|>image<|/ref|><|det|>[[145, 297, 283, 430]]<|/det|>  
+
+<|ref|>sub_title<|/ref|><|det|>[[140, 461, 750, 479]]<|/det|>
+## Is UCP1 expression more elevated in BAT from females than from males?  
+
+<|ref|>text<|/ref|><|det|>[[140, 483, 828, 542]]<|/det|>
+We measured the expression levels of Ucp1 in BAT. The expression levels of Ucp1 in BAT were higher in females than in males (Fig. S1a), consistent with the enhancement of BAT thermogenesis in female mice.  
+
+<|ref|>text<|/ref|><|det|>[[140, 546, 839, 586]]<|/det|>
+These results have been added to the Results section of the revised manuscript (Lines 88- 89: Supplementary Figure 1a).  
+
+<|ref|>sub_title<|/ref|><|det|>[[140, 610, 835, 648]]<|/det|>
+## In Figure 3 and in Figure 7, the expression of ChREBPa (RT-qPCR) as well as ChREBP protein content should be measured (actually in males and in females)  
+
+<|ref|>text<|/ref|><|det|>[[139, 652, 853, 800]]<|/det|>
+We measured Chrebpa gene expression levels and ChREBP protein levels in the BAT of both PGC- 1α KO mice and TMX- , an estrogen receptor antagonist, administered mice. Neither PGC- 1α KO (Fig. S3a) nor TMX administration (Fig. S7b) resulted in any changes in the expression levels of Chrebpa in either sex. Moreover, the expression levels of the ChREBP protein in BAT were higher in females than in males. In females, both PGC- 1α KO (Fig. S3b) and TMX administration (Fig. S7c) reduced BAT ChREBP protein expression levels, whereas in males, neither PGC- 1α KO nor TMX treatment induced such reductions.  
+
+<|ref|>text<|/ref|><|det|>[[140, 803, 848, 842]]<|/det|>
+These results have been added to the Results section of the revised manuscript (Lines 162- 166: Supplementary Figure 3a- b, and Lines 294- 296: Supplementary Figure 7b- c).  
+
+<|ref|>text<|/ref|><|det|>[[137, 867, 850, 885]]<|/det|>
+What are the functional consequences of the OVX- treatment on mitochondrial shape,
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[140, 120, 316, 136]]<|/det|>
+## /structure/function ?  
+
+<|ref|>text<|/ref|><|det|>[[139, 141, 852, 285]]<|/det|>
+In the BAT of ovariectomized mice, the mitochondrial area was smaller, and the total length of the cristae was also reduced compared with that in the sham- operated female mice (Fig. S7g- i). In addition, VO2 after NE administration was reduced in ovariectomized mice compared with that in sham- operated female mice (Fig. S7f). These results were similar to those observed in the TMX- administered female mice in this study. Overall, these results suggest that estrogen signaling plays an important role in BAT thermogenesis in female mice.  
+
+<|ref|>text<|/ref|><|det|>[[140, 290, 848, 329]]<|/det|>
+These results have been added to the Results section of the revised manuscript (Lines 302- 306: Supplementary Figure 7f- i).  
+
+<|ref|>sub_title<|/ref|><|det|>[[140, 354, 810, 393]]<|/det|>
+## FGF21, a key actor of BAT function, known to be regulated in a sexual dimorphic manner should be measured in all the experiments presented.  
+
+<|ref|>text<|/ref|><|det|>[[140, 397, 833, 500]]<|/det|>
+We measured the expression levels of Fgf21 in BAT in all experiments. No significant differences were observed in the expression levels of Fgf21 between the control and intervention groups as well as between males and females (Fig. S3c, and as shown in the figure below). These results suggest that FGF21 did not play an important role in the sex difference in BAT thermogenesis in this model.  
+
+<|ref|>text<|/ref|><|det|>[[140, 504, 848, 543]]<|/det|>
+These results have been added to the Results section of the revised manuscript (Lines 166- 168: Supplementary Figure 3c).  
+
+<|ref|>image<|/ref|><|det|>[[147, 572, 600, 710]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[140, 740, 238, 755]]<|/det|>
+Reviewer #3  
+
+<|ref|>text<|/ref|><|det|>[[139, 760, 849, 864]]<|/det|>
+However, tetra linoleoyl CL is the major functional cardiolipins, and it is NOT dependent on DNL, since linoleic acid is an essential acid and can only be taken up from dietary sources. It is recommended to show some data that DNL is in fact important in the proposed mechanism using isotope tracing experiments, such as 3H labeled glucose, or 3H H2O.  
+
+<|ref|>text<|/ref|><|det|>[[137, 867, 839, 885]]<|/det|>
+We deeply appreciate your valuable advice and suggestions. As suggested, we performed
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[137, 118, 856, 435]]<|/det|>
+lipidomics after \(\mathsf{D}_2\mathsf{O}\) administration and found that the labeled components of CL(18:2)4 were not detected in the PGC- 1α KO and Control mice of both sexes.. This result clearly indicates that PGC- 1α does not enhance CL(18:2)4 production through DNL in the BAT of female mice. In addition, we found that ChREBPβ knockdown did not affect CL(18:2)4 levels in female BAT (Fig. 6f). Overall, these results suggest that PGC- 1α increases CL(18:2)4 levels through molecular mechanism(s) other than ChREBPβ- mediated DNL in female BAT. In addition, we found that several molecular species of ether- linked phosphatidylethanolamine (PE), which were decreased in the female BAT of PGC- 1α KO mice (Fig. S5e), were commonly reduced following BAT ChREBPβ knockdown in female mice (Fig. 6g). The labeled components of ether- linked PE(O- 16:1_18:1) showed a significant decrease in lipidomics after \(\mathsf{D}_2\mathsf{O}\) administration in female Pgc- 1α KO mice (Fig. 6h). These results suggest that ether- linked PE is synthesized via PGC- 1α- ChREBPβ- mediated DNL. Recent studies have highlighted that ether- linked phospholipids play a crucial role in maintaining the morphology and function of the mitochondria (PMID: 37069167, 38129691).  
+
+<|ref|>text<|/ref|><|det|>[[140, 439, 830, 499]]<|/det|>
+In summary, PGC- 1α may play a pivotal role in the mitochondrial function of female BAT through both ChREBPβ- dependent de novo ether- linked PE production and ChREBPβ- independent CL(18:2)4 production.  
+
+<|ref|>text<|/ref|><|det|>[[140, 504, 855, 564]]<|/det|>
+These observations have been added to the Results section of the revised manuscript (Lines 263- 272: Figure 6f- h, Supplementary Figure 5e) and discussed in the Discussion section of the revised manuscript (Lines 370- 373 and Lines 379- 386).  
+
+<|ref|>sub_title<|/ref|><|det|>[[141, 590, 411, 605]]<|/det|>
+## Below are some minor concerns:  
+
+<|ref|>sub_title<|/ref|><|det|>[[140, 610, 653, 627]]<|/det|>
+## 1. Please indicate the weights/mass of BAT in different mice.  
+
+<|ref|>text<|/ref|><|det|>[[140, 632, 802, 692]]<|/det|>
+We measured the BAT mass in PGC- 1α KO or TMX-, an estrogen receptor antagonist, administered mice. As shown in the figure below, neither PGC- 1α KO nor TMX administration resulted in any change in BAT weight in either sex.  
+
+<|ref|>image<|/ref|><|det|>[[147, 709, 660, 850]]<|/det|>
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[139, 140, 794, 200]]<|/det|>
+## 2. Figure 1d and 1f, Male KO mice also showed similar trends with female mice although it is not statistically significant. Is it because of the larger individual variation?  
+
+<|ref|>text<|/ref|><|det|>[[139, 204, 852, 393]]<|/det|>
+We agree that body temperature during cold exposure tended to be lower in Male KO mice than in Male Control mice. However, as the reviewer pointed out, these differences were not statistically significant. Moreover, the effects of PGC- 1α KO on BAT, such as energy expenditure (Figure 1g), gene expression (Figure 3a- c), metabolite profile (Figure 4a- e), lipid profile (Figure 5a- i), and mitochondrial structure (Figure 2a- c) and activity (Figure 2d and Figure 4e), were all observed only in female BAT, which is completely consistent with the fact that cold tolerance was significantly reduced only in female mice (Figure 1d, f). Therefore, the lack of a significant reduction in cold tolerance in male mice may be due to sex differences in BAT function.  
+
+<|ref|>sub_title<|/ref|><|det|>[[140, 418, 758, 435]]<|/det|>
+## 3. How to explain the difference of GTT in male and female mice in Fig.s1f  
+
+<|ref|>text<|/ref|><|det|>[[139, 440, 848, 544]]<|/det|>
+A previous study (PMID: 22645355) reported that male adipocyte- specific PGC- 1α KO mice were systemic glucose intolerant. Our results are compatible with those of this previous report. Given that male mice have poorer glucose tolerance than female mice, impaired glucose tolerance may be more pronounced in male mice than in female mice, although further studies are required to clarify the mechanisms underlying this sex difference.  
+
+<|ref|>sub_title<|/ref|><|det|>[[140, 568, 835, 606]]<|/det|>
+## 4. IN Figure 4A, please justify why they only focused on polar metabolites using CEMS methods. IN the heatmap, how do they normalize the different metabolites?  
+
+<|ref|>text<|/ref|><|det|>[[139, 610, 844, 778]]<|/det|>
+We focused only on polar metabolites because most molecules produced when the mitochondrial TCA cycle and electron transport chain are activated during BAT thermogenesis are water- soluble and easily ionized molecules. CE- MS methods were used because they are suitable for analyzing polar metabolites. In the heatmap analysis, we used the following formula for normalization: \(Z = (X - \mu) / \sigma\) (Z: standardized value of each metabolite, X: Relative area of each metabolite, \(\mu\) : mean of each metabolite, \(\sigma\) : SD of each metabolite)  
+
+<|ref|>sub_title<|/ref|><|det|>[[140, 803, 848, 821]]<|/det|>
+## 5. In Figure 5e and 5f, why the patterns of CL(18:2) are different in these two figures?  
+
+<|ref|>text<|/ref|><|det|>[[140, 825, 856, 884]]<|/det|>
+Figure 5e shows the proportion of each CL molecular species, including CL(18:2)4, relative to the total lipid content in BAT. In contrast, Figure 5f presents the proportion of the CL(18:2)4 molecular species normalized to the total CL content in BAT. The proportions of CL among
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[139, 118, 858, 244]]<|/det|>
+all lipids are very small, and those of individual CL species are greatly influenced not only by changes in the absolute amount of individual CL species but also by changes in the absolute amount of total lipids excluding CL. Therefore, to evaluate the changes in CL(18:2)4 within CL, we measured the proportion of CL(18:2)4 among CL as previously reported (PMID: 16855048) and found that CL(18:2)4 was actually reduced only in female knockout mice, as shown in Figure 5f.  
+
+<|ref|>sub_title<|/ref|><|det|>[[140, 269, 722, 286]]<|/det|>
+## 6. In line 644, what does "10 uL brown adipose tissue samples" mean?  
+
+<|ref|>text<|/ref|><|det|>[[140, 291, 844, 328]]<|/det|>
+We thank the reviewer for this comment. This is a typographical error; the correct notation should be "1 mg brown adipose tissue." We have corrected this error in the text.
+
+<--- Page Split --->

peer_reviews/123bd0ee09c9429721618645aa9aeb915f306f1aae5c4eccbeaf12f98f492aec/supplementary_0_Peer Review File/images_list.json

@@ -0,0 +1,10 @@
+[
+  {
+    "type": "image",
+    "img_path": "images/Figure_2.jpg",
+    "caption": "From Fig. 2. (A-E) Acute nociceptive effects of \\(\\mathsf{PGE}_2\\) in IL6-treated mice. Note the persistence of the effects at the 1-h time point even in mice treated with NAAA inhibitor ARN19702 (please see Fig. 2 for details). #P < 0.05, #P < 0.01, and #P < 0.001 compared to veh/veh (n = 10). See Figure S9 for details.",
+    "footnote": [],
+    "bbox": [],
+    "page_idx": 0
+  }
+]

peer_reviews/123bd0ee09c9429721618645aa9aeb915f306f1aae5c4eccbeaf12f98f492aec/supplementary_0_Peer Review File/supplementary_0_Peer Review File.mmd

@@ -0,0 +1,283 @@
+
+# nature portfolio  
+
+Peer Review File  
+
+NAAA- regulated lipid signaling in monocytes controls the induction of hyperalgesic priming in mice  
+
+![](images/Figure_2.jpg)
+  
+
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+
+Reviewer #1 (Remarks to the Author):  
+
+N- acylethanolamine- hydrolyzing acid amidase (NAAA) is a lysosomal enzyme abundant in monocytes and macrophages. NAAA hydrolyzes N- acylethanolamines, including palmitoylethanolamide (PEA) and oleoylethanolamide (OEA). Over the years, the authors have studied the anti- inflammatory and analgesic actions of N- acylethanolamines in vivo and have developed specific NAAA inhibitors which exhibit anti- inflammatory and analgesic effects. In the present study the authors extended their interests to the involvement of NAAA in hyperalgesic priming. They used specific NAAA inhibitors and plural kinds of NAAA- deficient mice to provide the evidence that NAAA- regulated signaling at PPAR- alpha directs monocytes and macrophages to initiate hyperalgesic priming in mice exposed to an inflammatory stimulus. This is the first report clearly showing that NAAA- regulated signaling is involved in hyperalgesic priming. Thus, the work is of significance to the field and rich in originality. The work sufficiently supports the conclusions. There are no flaws in the data analysis, interpretation and conclusions. The methodology is also sound. I think that the work meets the expected standards in our field.  
+
+I just have some minor comments.  
+
+1. In Fig. 4E, the symbols for significance should be placed near the line of CD11b-/-.  
+
+2. In Fig. 4F1, the asterisks should be explained in the legend.  
+
+3. In Fig. 4G1, the symbols may be asterisks.  
+
+Reviewer #2 (Remarks to the Author):  
+
+This study shows that the induction of hyperalgesic priming by intraplantar IL- 6 to intraplantar PGE2 involves monocytes and macrophages. These immune cells are the source for hydrolase NAAA that acts on two lipids (PEA and OEA) which are agonists for intracellular PPAR- alpha signalling. The interruption of this pathway in monocytes and macrophages by a peripheralized NAAA inhibitor blocks the induction of IL- 6- mediated hyperalgesic priming in male and female mice.  
+
+In addition, HP induction is impaired in global NAAA knockout and following silencing of NAAA in CD11b cells and by clotronate liposome and CSF1 receptor antagonist treatments.  
+
+These data suggest that monocyte/macrophages are critical for the induction of hyperalgesic priming and in such cells lipid signalling regulated by NAAA plays a mechanistic role.  
+
+This is an interesting possibility for a novel neuro- immune interaction- mediated mechanisms in the induction of persistent pain. However, the current behavioural data in transgenic mice should be paired with characterization of monocytes/macrophages in the injected paw before and after intraplantar injections of IL- 6 and PGE2.
+
+<--- Page Split --->
+
+More specifically some of the following points should be addressed:  
+
+1. Do hind paw macrophages express and regulate expression of IL-6 receptor? What levels of NAAA, PEA, OEA and palmitic acid do they contain before, 6 hours and 3 days after IL-6 injections?  
+
+2. What is the level of NAAA, PEA, OEA and PA in macrophages isolated from the hind paw of NAAACD11b-/- and PparaCD11b-/- mice?  
+
+3. Which phenotypes do monocytes and macrophages acquire at 6 h and 3 days after intrapalantar IL-6 and after PGE2? Are they more likely to release pro-nociceptive chemicals? (Figure 6).  
+
+4. How do the authors explain IL-6 and PGE2 pro-nociceptive effects at 1 h after injections in NAAACD11b-/- in figure 3?  
+
+Reviewer #3 (Remarks to the Author):  
+
+In this manuscript, the authors aim to identify molecular mechanism of acute to chronic pain transition. They concluded a critical role of NAAA on monocyte- derived cells and PPAR- a receptors in the event by using a mouse model of hyperalgesic priming. The study is carefully designed, with various experimental approaches. While a large amount of data supported most of their conclusion, I have several major concerns on the limitation of the study.  
+
+- Model of hyperalgesic priming: I greatly appreciate the model provides us an evidence that prior priming nociceptors could lead increased sensitivity to subsequent stimulation. The authors in this study demonstrated potential underlying mechanism mediated by monocyte-derived cells-associated NAAA. However, this is a very restricted model, only limited to inflammatory triggers, with specific agents and specific doses. I am sure not if the agent or the dose changed, whether the 72hour-interval is still valid, not saying if this is valid with a non-inflammatory trigger. It may mislead knowledge users by generalize the findings.  
+
+- Pain behavioral testing methods: Only heat sensitivity was assessed in the study. Whether such priming is also effective in mechanical and cold sensitivity is unknown. Whether NAAA mediates other pain modalities following the priming was not assessed either in the study.  
+
+- Skin macrophages: Two experimental approaches to deplete monocytes used in the study are indeed not ideal. Both are more effective and suitable in depleting macrophages/microglia than monocytes. The authors repeatedly indicated that observed effects are derived from monocytes/macrophages. They
+
+<--- Page Split --->
+
+excluded CNS microglia effect, but they didn't touch anything on skin macrophages, which is essential. If we believe there is a crucial peripheral contribution of NAAA in inflammatory mediator induced priming, detailed changes of macrophages in affected paw are indispensable, as they are direct players. Such data may also help to understand why it works not before or not after 72hr interval.  
+
+- Clinical implications: It is difficult for me to imagine in what clinical setting where the findings from the current study can apply. What does this 72 hr interval, not before or not after, mean for patients who come to see a physician for an acute pain?  
+
+Thus, in general, I recognize the value and the quality of the study in identifying molecular mechanism of hypersensitivity priming in a specific setting, I feel the conclusion, including the title of the study was overstated. Chronic pain is complex, I believe we need to develop precision medicine for each specific type of chronic pain.  
+
+Reviewer #4 (Remarks to the Author):  
+
+The manuscript "NAAA- regulated lipid signaling in monocyte- derived cells controls the induction of hyperalgesic priming", by Fotio et al, provides additional mechanistic information regarding a model of the transition from acute to chronic pain, hyperalgesic priming (HP). This is an interesting study, with the experiments appropriately conducted, clearly exemplifying the involvement of the immune system in the development of persistent nociceptive sensitization. The conclusions certainly add to the characterization of such phenomenon, previously shown to play a role in the persistence of pain observed in some clinical conditions. Although I do not see a reason preventing its publication, I would appreciate if the authors could clarify or include some minor comments in the final version of the manuscript:  
+
+- Experimentally, HP is defined by the potentiation (prolongation) of the mechanical hyperalgesia induced by PGE2 (>4h) in the paw when compared to a non-primed paw (<2h), indicating the hyper-responsive state triggered by the inflammation in the primary neuron. In the current study the authors evaluate the effect of PGE2 on the thermal nociceptive threshold in primed paws. What is the time course of the response produced by PGE2 on the thermal sensitivity in a normal, non-primed paw? Including this information would help the reader to visualize the increased sensitivity of the nociceptor produced by the priming stimulus, for comparison.  
+
+- Induction of HP by carrageenan is, in rats, sexually dimorphic, modulated by estrogen at the level of PKCepsilon. In mice, HP is also sexually dimorphic. However, instead of being an "all or nothing" phenomenon as in rats, this difference in mice seems to be more complex. Can the authors include a comment about these differences in the Discussion? Since the development of HP in mice of both sexes is suggested to occur at the level of NAAA, the upstream differential mechanism could be mentioned (or speculated), specifically considering that NAAA is important only during the incubation phase, when protein translation is happening.  
+
+- Why did the authors use the tail flick method to determine nociceptive thresholds if all the relevant behavior experiments were performed using the Hargreaves plantar test in the paw?
+
+<--- Page Split --->
+
+- Were the compounds injected in the paw intradermally or subcutaneously?  
+
+The information provided by this study contribute to the knowledge about the mechanisms involved in the transition from acute to chronic pain, which is of utmost importance. I look forward to receiving the authors' response and comments added to the revised version of this article.
+
+<--- Page Split --->
+
+## RESPONSE TO REVIEWER COMMENTS  
+
+We are grateful to the Reviewers for their constructive and stimulating comments. We addressed them with several new experiments and substantive text edits, highlighted in red in this resubmission.  
+
+## Reviewer #1  
+
+N- acylethanolamine- hydrolyzing acid amidase (NAAA) is a lysosomal enzyme abundant in monocytes and macrophages. NAAA hydrolyzes N- acylethanolamines, including palmitoylethanolamide (PEA) and oleoylethanolamide (OEA). Over the years, the authors have studied the anti- inflammatory and analgesic actions of N- acylethanolamines in vivo and have developed specific NAAA inhibitors which exhibit anti- inflammatory and analgesic effects. In the present study the authors extended their interests to the involvement of NAAA in hyperalgesic priming. They used specific NAAA inhibitors and plural kinds of NAAA- deficient mice to provide the evidence that NAAA- regulated signaling at PPAR- alpha directs monocytes and macrophages to initiate hyperalgesic priming in mice exposed to an inflammatory stimulus. This is the first report clearly showing that NAAA- regulated signaling is involved in hyperalgesic priming. Thus, the work is of significance to the field and rich in originality. The work sufficiently supports the conclusions. There are no flaws in the data analysis, interpretation and conclusions. The methodology is also sound. I think that the work meets the expected standards in our field. I just have some minor comments.  
+
+We are grateful for this positive evaluation of our work.  
+
+1. In Fig. 4E, the symbols for significance should be placed near the line of CD11b-/-.  
+
+2. In Fig. 4F1, the asterisks should be explained in the legend.  
+
+3. In Fig. 4G1, the symbols may be asterisks.  
+
+Done.  
+
+## Reviewer #2  
+
+This study shows that the induction of hyperalgesic priming by intraplantar IL- 6 to intraplantar PGE2 involves monocytes and macrophages. These immune cells are the source for hydrolase NAAA that acts on two lipids (PEA and OEA) which are agonists for intracellular PPAR- alpha signalling. The interruption of this pathway in monocytes and macrophages by a peripheralized NAAA inhibitor blocks the induction of IL- 6- mediated hyperalgesic priming in male and female mice. In addition, HP induction is impaired in global NAAA knockout and following silencing of NAAA in CD11b cells and by clodronate liposome and CSF1 receptor antagonist treatments.  
+
+These data suggest that monocyte/macrophages are critical for the induction of hyperalgesic priming and in such cells lipid signalling regulated by NAAA plays a mechanistic role. This is an interesting possibility for a novel neuro- immune interaction- mediated mechanisms in the induction of persistent pain. However, the current behavioural data in transgenic mice should be paired with characterization of monocytes/macrophages in the injected paw before and after
+
+<--- Page Split --->
+
+intraplantar injections of IL- 6 and PGE2. More specifically some of the following points should be addressed: 1. Do hind paw macrophages express and regulate expression of IL- 6 receptor? What levels of NAAA, PEA, OEA and palmitic acid do they contain before, 6 hours and 3 days after IL- 6 injections? 2. What is the level of NAAA, PEA, OEA and PA in macrophages isolated from the hind paw of NAAACD11b-/- and PparaCD11b-/- mice?  
+
+We thank the Reviewer for this important comment, which was echoed by a similar one made by Reviewer 3. Prompted by these suggestions, we conducted a new experiment to assess whether resident macrophages contribute to the initiation of hyperalgesic priming. We removed macrophages from the hind paws of mice with two intraplantar injections of clodronate liposomes (PBS liposomes were injected as control) and examined the effects of IL- 6 administration in the same paws. The results, which are reported in new Figure S9, show that clodronate treatment substantially reduced the number of \(\mathrm{CD68^{+}}\) macrophages in paw tissue (Fig. S9B) but did not prevent IL- 6- induced hyperalgesic priming (new Fig. S9E).  
+
+![PLACEHOLDER_6_0]
+  
+
+From Fig. S9. (B) Effect of PBS- containing liposomes (left) and clodronate- containing liposomes (right) on \(\mathrm{CD68 + }\) cells (resident macrophages) in mouse paw skin. (C) Effect of PBS- or clodronate- containing liposomes on the induction of hyperalgesic priming (response to PGE2 six days after IL- 6). See Figure S9 for additional details.  
+
+By contrast, clodronate delayed the acute response to IL- 6, suggesting that local macrophages contribute in part to such response.  
+
+![PLACEHOLDER_6_1]
+  
+
+From Fig. S9. (E) Effect of PBS- or clodronate- containing liposomes on the acute nociceptive response to IL- 6. # \(P < 0.05\) compared to PBS ( \(n = 8\) ). See Figure S9 for details.  
+
+Along with the data shown in Figure 5 (effects of systemic clodronate and PLX- 5622) these findings indicate that blood- borne monocytes, rather than resident macrophages, are required for priming initiation. We modified our hypothetical model (and the title of the study) to reflect these new results (please see new Fig. 7).
+
+<--- Page Split --->
+
+Which phenotypes do monocytes and macrophages acquire at 6 h and 3 days after intraplantar IL- 6 and after PGE2? Are they more likely to release pro- nociceptive chemicals? (Figure 6).  
+
+Based on our results showing that monocytes, not resident macrophages, are involved in priming, we addressed this question by examining the molecular phenotype of circulating immune cells by high- resolution mass cytometry by time of flight (CyTOF). Mice received intraplantar IL- 6 injections and, 72 hours later (i.e., at the tail end of the incubation phase of priming), cardiac blood was collected for CyTOF analyses. The results, which are illustrated in new Figure 6, indicate that the incubation phase of hyperalgesic priming coincides with the emergence of one or more subpopulation(s) of activated circulating monocytes. For example:  
+
+![PLACEHOLDER_7_0]
+  
+
+From Fig. 6. (B) Optimized Stochastic Neighbor Embedding (opt- SNE) plots depicting density heatmaps of cells expressing the markers of monocyte activation CCR2 (top), CD43 (middle), and CX3CR1 (bottom) in vehicle- (left) and IL- 6- treated (right) mice. See new Figure 6 for details.  
+
+4. How do the authors explain IL-6 and PGE2 pro-nociceptive effects at 1 h after injections in NAAACD11b-/- in figure 3?  
+
+As pointed out by the Reviewer, Figure 3E shows that, one h after injection, PGE2 exerts a nociceptive effect in both control Naaa f/l and Naaa CD11b-/- mice:  
+
+![PLACEHOLDER_7_1]
+  
+
+From Fig. 3. (E) Acute nociceptive effects of PGE2 in Naaa CD11b-/- mice (green symbols) and control Naaa f/l mice (magenta symbols) #P < 0.001 (n = 8). See Figure 3 for details.
+
+<--- Page Split --->
+
+This effect is expected because the 1h time point measures the acute nociceptive response to \(\mathsf{PGE}_2\) , which is generally not affected by pharmacological or genetic interventions targeting NAAA. See, for example:  
+
+![PLACEHOLDER_8_0]
+
+<center>From Fig. 2. (A-E) Acute nociceptive effects of \(\mathsf{PGE}_2\) in IL6-treated mice. Note the persistence of the effects at the 1-h time point even in mice treated with NAAA inhibitor ARN19702 (please see Fig. 2 for details). #P < 0.05, #P < 0.01, and #P < 0.001 compared to veh/veh (n = 10). See Figure S9 for details. </center>  
+
+One exception is represented by global Naaa ko mice, which exhibited a weaker response to \(\mathsf{PGE}_2\) at the 1- h time point (Fig. 3). This could be due to NAAA deletion in as- yet- unidentified cells that are not readily accessible to the NAAA inhibitors used in our study.  
+
+## Reviewer #3  
+
+In this manuscript, the authors aim to identify molecular mechanism of acute to chronic pain transition. They concluded a critical role of NAAA on monocyte- derived cells and PPAR- a receptors in the event by using a mouse model of hyperalgesic priming. The study is carefully designed, with various experimental approaches. While a large amount of data supported most of their conclusion, I have several major concerns on the limitation of the study.  
+
+Model of hyperalgesic priming: I greatly appreciate the model provides us an evidence that prior priming nociceptors could lead increased sensitivity to subsequent stimulation. The authors in this study demonstrated potential underlying mechanism mediated by monocyte- derived cells- associated NAAA. However, this is a very restricted model, only limited to inflammatory triggers, with specific agents and specific doses. I am sure not if the agent or the dose changed, whether the 72hour- interval is still valid, not saying if this is valid with a non- inflammatory trigger. It may mislead knowledge users by generalize the findings.  
+
+Hyperalgesic priming can be induced by a variety of stimuli, including proinflammatory cytokines (e.g., IL- 6, TNF- \(\alpha\) ), growth factors (e.g., NGF, CSF- 1), inflammatory triggers (e.g., carrageenan), and tissue damage (e.g., surgical paw incision). Its clinical relevance has been convincingly argued by various investigators (Levine, Price, etc.) who pointed out that "... the experimental framework of the hyperalgesic priming model provides important insight into clinical chronic pain because it captures the recurrent nature of some of the most common pathological pain conditions." (Kandasamy R, Price TJ. The pharmacology of nociceptor priming. Handb Exp Pharmacol. 2015; 227:15- 37. PMID: 25846612) Our previous submission described the results of experiments with three distinct proinflammatory stimuli – IL- 6, TNF- \(\alpha\) , and carrageenan. Nevertheless, to address the spirit of the Reviewer's concern, we have included new data using a fourth stimulus, paw incision, which has a strong nociceptive component. The results, reported in new Figure S7, show
+
+<--- Page Split --->
+
+that administration of a NAAA inhibitor in the first 72h after the incision fully prevents priming initiation, lending further support to the general validity of our findings.  
+
+Pain behavioral testing methods: Only heat sensitivity was assessed in the study. Whether such priming is also effective in mechanical and cold sensitivity is unknown. Whether NAAA mediates other pain modalities following the priming was not assessed either in the study.  
+
+We assessed both heat and mechanical hypersensitivity, with similar results. Representative sets of mechanical hypersensitivity data are included in new Figure S2 and S4.  
+
+Skin macrophages: Two experimental approaches to deplete monocytes used in the study are indeed not ideal. Both are more effective and suitable in depleting macrophages/microglia than monocytes.  
+
+We direct the Reviewer's attention to the results reported in Figure 5, which show that both approaches (clodronate and PLX- 5622) substantially reduced monocyte numbers in the bloodstream.  
+
+The authors repeatedly indicated that observed effects are derived from monocytes/macrophages. They excluded CNS microglia effect, but they didn't touch anything on skin macrophages, which is essential. If we believe there is a crucial peripheral contribution of NAAA in inflammatory mediator induced priming, detailed changes of macrophages in affected paw are indispensable, as they are direct players. Such data may also help to understand why it works not before or not after 72hr interval.  
+
+We thank the Reviewer for this suggestion, which we have addressed with the new experiment described in our response to Reviewer 2. Please see above.  
+
+Clinical implications: It is difficult for me to imagine in what clinical setting where the findings from the current study can apply. What does this 72 hr interval, not before or not after, mean for patients who come to see a physician for an acute pain?  
+
+This basic science study is not meant to have immediate repercussions on clinical practice. It does have, however, potential clinical significance in that it suggests that postsurgical NAAA inhibition (but not postsurgical treatment with standard analgesics) may prevent the transition to pain chronicity, a significant problem for patients who undergo thoracotomy, knee arthroplasty, mastectomy, and other invasive surgeries.  
+
+Thus, in general, I recognize the value and the quality of the study in identifying molecular mechanism of hypersensitivity priming in a specific setting, I feel the conclusion, including the title of the study was overstated. Chronic pain is complex, I believe we need to develop precision medicine for each specific type of chronic pain.  
+
+Our views about the complexity of chronic pain are very similar to the Reviewer's, as affirmed in the Discussion section of the manuscript:  
+
+"Chronic pain states are widely heterogeneous in causes, symptoms, impact on function, and temporal development<sup>3</sup>. This diversity justifies skepticism toward a simplistic view of the progression to pain chronicity as transformation of one mechanistic type of pain (e.g., acute pain associated with injury) into another (e.g., neuropathic pain)<sup>99</sup>."
+
+<--- Page Split --->
+
+Thus, we are surprised by the Reviewer's statement that "the conclusion, including the title of the study was overstated." The title of our manuscript (NAAA- regulated lipid signaling in monocytes controls the induction of hyperalgesic priming) accurately reflects the findings presented. Our conclusion is equally cautious and, we believe, fully warranted by the data: "Clinical studies should determine whether algostatic agents might offer a strategy to prevent chronic pain after invasive surgery and other kinds of physical trauma." However, when interpretation could have been ambiguous, we replaced 'chronic pain' with 'persistent pathological nociception' (see, for example, Abstract, line 38).  
+
+## Reviewer #4  
+
+The manuscript "NAAA- regulated lipid signaling in monocyte- derived cells controls the induction of hyperalgesic priming", by Fotio et al, provides additional mechanical information regarding a model of the transition from acute to chronic pain, hyperalgesic priming (HP). This is an interesting study, with the experiments appropriately conducted, clearly exemplifying the involvement of the immune system in the development of persistent nociceptive sensitization. The conclusions certainly add to the characterization of such phenomenon, previously shown to play a role in the persistence of pain observed in some clinical conditions.  
+
+We appreciate the positive comments.  
+
+Although I do not see a reason preventing its publication, I would appreciate if the authors could clarify or include some minor comments in the final version of the manuscript:  
+
+Experimentally, HP is defined by the potentiation (prolongation) of the mechanical hyperalgesia induced by PGE2 (>4h) in the paw when compared to a non- primed paw (<2h), indicating the hyper- responsive state triggered by the inflammation in the primary neuron. In the current study the authors evaluate the effect of PGE2 on the thermal nociceptive threshold in primed paws. What is the time course of the response produced by PGE2 on the thermal sensitivity in a normal, non- primed paw? Including this information would help the reader to visualize the increased sensitivity of the nociceptor produced by the priming stimulus, for comparison.  
+
+We have added a new supplemental figure (Fig. S2) that contains the requested information.  
+
+Induction of HP by carrageenan is, in rats, sexually dimorphic, modulated by estrogen at the level of PKCepsilon. In mice, HP is also sexually dimorphic. However, instead of being an "all or nothing" phenomenon as in rats, this difference in mice seems to be more complex. Can the authors include a comment about these differences in the Discussion? Since the development of HP in mice of both sexes is suggested to occur at the level of NAAA, the upstream differential mechanism could be mentioned (or speculated), specifically considering that NAAA is important only during the incubation phase, when protein translation is happening.  
+
+We do not know how inflammatory challenges trigger NAAA activation and whether this process might be sexually dimorphic. But we appreciate the point raised by the Reviewer and addressed its spirit by adding the following comment to the Discussion section of the manuscript (p. 17):  
+
+"In this context, an important question that remains to be answered pertains to the mechanisms through which inflammatory stimuli lead to the suppression of NAPE- PLD
+
+<--- Page Split --->
+
+transcription and the enhancement of NAAA activity. Elucidating such mechanisms might shed light on aspects of HP – such as the existence of sexual dimorphisms in rats<sup>88- 90</sup> and mice<sup>91</sup> – which are presently unclear."  
+
+Why did the authors use the tail flick method to determine nociceptive thresholds if all the relevant behavior experiments were performed using the Hargreaves plantar test in the paw?  
+
+We routinely use the tail- flick method to identify potential alterations in baseline nociceptive threshold in newly developed mouse lines such as those introduced in the present study. Here, we combined tail- flick (Fig. S9) and Hargreaves plantar tests (Fig. 3D and Fig. 4G1) to obtain a more complete view of the lines' nociceptive phenotype.  
+
+Were the compounds injected in the paw intradermally or subcutaneously?  
+
+Compounds were injected subcutaneously, between skin and muscular fascia/tendon, as customary in this model. We clarified this point in the Materials and Methods section of the manuscript.  
+
+The information provided by this study contribute to the knowledge about the mechanisms involved in the transition from acute to chronic pain, which is of utmost importance. I look forward to receiving the authors' response and comments added to the revised version of this article.  
+
+Thank you.
+
+<--- Page Split --->
+
+Reviewer #1 (Remarks to the Author):  
+
+The authors appropriately revised the manuscript in response to my comments.  
+
+Reviewer #2 (Remarks to the Author):  
+
+My points have been partially addressed.  
+
+Specifically, whilst I suggested to characterise monocytes/macrophages at the site of IL- 6 injection (hind paw) the authors have looked at blood monocytes. They provide evidence that intraplantar injection of IL- 6 is associated with more CCR2+ monocytes circulating in blood. However, classical monocytes will have to infiltrate tissue to come in the vicinity of primary afferent terminals and promote nociceptive signalling.  
+
+Whether CCR2+ monocytes infiltrate the site of injection remains to be established and Figure 7 schematic should indicate that this study does not provide direct evidence that "monocytes migrate to target tissue".  
+
+New Figure 6: How do the authors explain presence of F4/80 positive cells (macrophages) in blood?  
+
+Discussion: It is possible that CCL2 is upregulated in endothelial cells, and sensory neurons and accumulates in primary afferent terminals. However, the authors provide no direct evidence that this happens in their experimental conditions.  
+
+Other points. Ref 12 line 57: This is an early observation and more recent evidence indicates no infiltration of monocytes in the spinal cord following peripheral nerve injury. The authors could rephrase their statement.  
+
+Line 167: replace "deleted" with "depleted".  
+
+Reviewer #3 (Remarks to the Author):  
+
+The authors addressed the reviewer's concerns. I don't have further comments.  
+
+Reviewer #4 (Remarks to the Author):  
+
+I am satisfied with the revision provided by the authors. I recommend this study for publication.
+
+<--- Page Split --->
+
+Specifically, whilst I suggested to characterise monocytes/macrophages at the site of IL- 6 injection (hind paw) the authors have looked at blood monocytes. They provide evidence that intraplantar injection of IL- 6 is associated with more CCR2+ monocytes circulating in blood. However, classical monocytes will have to infiltrate tissue to come in the vicinity of primary afferent terminals and promote nociceptive signalling. Whether CCR2+ monocytes infiltrate the site of injection remains to be established and Figure 7 schematic should indicate that this study does not provide direct evidence that "monocytes migrate to target tissue".  
+
+We clarified that Figure 7 provides a hypothetical model which summarizes the data presented in our study and suggests a roadmap for future experiments. Please see Figure 7 legend.  
+
+New Figure 6: How do the authors explain presence of F4/80 positive cells (macrophages) in blood?  
+
+Though unlikely, it is possible that the antibody we used cross- reacted with a different monocyte antigen. Owing to this uncertainty, we removed F4/80 data from Figure 6C.  
+
+Discussion: It is possible that CCL2 is upregulated in endothelial cells, and sensory neurons and accumulates in primary afferent terminals. However, the authors provide no direct evidence that this happens in their experimental conditions.  
+
+There is substantial evidence that peripheral nerve injury upregulates CCL2 in DRG neurons, but future experiments will have to address this question as it pertains to the model used here.  
+
+Other points. Ref 12 line 57: This is an early observation and more recent evidence indicates no infiltration of monocytes in the spinal cord following peripheral nerve injury. The authors could rephrase their statement.  
+
+Evidence indicates that spinal nerve resection does not stimulate monocytes migration toward the spinal cord (PMID 27373153, presumably the study the reviewer had in mind). The hyperalgesic priming model we used, however, is quite different and, possibly, mechanistically closer to the one used in Ref 12 (partial sciatic nerve ligation). Indeed, both hyperalgesic priming and sciatic nerve ligation exhibit a rather strong inflammatory component that is not observed after spinal nerve resection.  
+
+Line 167: replace "deleted" with "depleted".  
+
+Done.
+
+<--- Page Split --->

peer_reviews/123bd0ee09c9429721618645aa9aeb915f306f1aae5c4eccbeaf12f98f492aec/supplementary_0_Peer Review File/supplementary_0_Peer Review File_det.mmd

@@ -0,0 +1,382 @@
+<|ref|>title<|/ref|><|det|>[[61, 40, 508, 90]]<|/det|>
+# nature portfolio  
+
+<|ref|>text<|/ref|><|det|>[[70, 110, 362, 139]]<|/det|>
+Peer Review File  
+
+<|ref|>text<|/ref|><|det|>[[70, 154, 852, 212]]<|/det|>
+NAAA- regulated lipid signaling in monocytes controls the induction of hyperalgesic priming in mice  
+
+<|ref|>image<|/ref|><|det|>[[57, 732, 240, 780]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[250, 732, 911, 785]]<|/det|>
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[116, 147, 392, 163]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[114, 175, 884, 394]]<|/det|>
+N- acylethanolamine- hydrolyzing acid amidase (NAAA) is a lysosomal enzyme abundant in monocytes and macrophages. NAAA hydrolyzes N- acylethanolamines, including palmitoylethanolamide (PEA) and oleoylethanolamide (OEA). Over the years, the authors have studied the anti- inflammatory and analgesic actions of N- acylethanolamines in vivo and have developed specific NAAA inhibitors which exhibit anti- inflammatory and analgesic effects. In the present study the authors extended their interests to the involvement of NAAA in hyperalgesic priming. They used specific NAAA inhibitors and plural kinds of NAAA- deficient mice to provide the evidence that NAAA- regulated signaling at PPAR- alpha directs monocytes and macrophages to initiate hyperalgesic priming in mice exposed to an inflammatory stimulus. This is the first report clearly showing that NAAA- regulated signaling is involved in hyperalgesic priming. Thus, the work is of significance to the field and rich in originality. The work sufficiently supports the conclusions. There are no flaws in the data analysis, interpretation and conclusions. The methodology is also sound. I think that the work meets the expected standards in our field.  
+
+<|ref|>text<|/ref|><|det|>[[116, 405, 367, 421]]<|/det|>
+I just have some minor comments.  
+
+<|ref|>text<|/ref|><|det|>[[114, 460, 728, 479]]<|/det|>
+1. In Fig. 4E, the symbols for significance should be placed near the line of CD11b-/-.  
+
+<|ref|>text<|/ref|><|det|>[[114, 490, 562, 506]]<|/det|>
+2. In Fig. 4F1, the asterisks should be explained in the legend.  
+
+<|ref|>text<|/ref|><|det|>[[115, 518, 437, 534]]<|/det|>
+3. In Fig. 4G1, the symbols may be asterisks.  
+
+<|ref|>text<|/ref|><|det|>[[116, 574, 392, 590]]<|/det|>
+Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[114, 602, 872, 694]]<|/det|>
+This study shows that the induction of hyperalgesic priming by intraplantar IL- 6 to intraplantar PGE2 involves monocytes and macrophages. These immune cells are the source for hydrolase NAAA that acts on two lipids (PEA and OEA) which are agonists for intracellular PPAR- alpha signalling. The interruption of this pathway in monocytes and macrophages by a peripheralized NAAA inhibitor blocks the induction of IL- 6- mediated hyperalgesic priming in male and female mice.  
+
+<|ref|>text<|/ref|><|det|>[[114, 704, 872, 740]]<|/det|>
+In addition, HP induction is impaired in global NAAA knockout and following silencing of NAAA in CD11b cells and by clotronate liposome and CSF1 receptor antagonist treatments.  
+
+<|ref|>text<|/ref|><|det|>[[114, 779, 856, 815]]<|/det|>
+These data suggest that monocyte/macrophages are critical for the induction of hyperalgesic priming and in such cells lipid signalling regulated by NAAA plays a mechanistic role.  
+
+<|ref|>text<|/ref|><|det|>[[114, 825, 870, 898]]<|/det|>
+This is an interesting possibility for a novel neuro- immune interaction- mediated mechanisms in the induction of persistent pain. However, the current behavioural data in transgenic mice should be paired with characterization of monocytes/macrophages in the injected paw before and after intraplantar injections of IL- 6 and PGE2.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 118, 611, 135]]<|/det|>
+More specifically some of the following points should be addressed:  
+
+<|ref|>text<|/ref|><|det|>[[115, 174, 852, 211]]<|/det|>
+1. Do hind paw macrophages express and regulate expression of IL-6 receptor? What levels of NAAA, PEA, OEA and palmitic acid do they contain before, 6 hours and 3 days after IL-6 injections?  
+
+<|ref|>text<|/ref|><|det|>[[115, 250, 789, 286]]<|/det|>
+2. What is the level of NAAA, PEA, OEA and PA in macrophages isolated from the hind paw of NAAACD11b-/- and PparaCD11b-/- mice?  
+
+<|ref|>text<|/ref|><|det|>[[115, 324, 864, 361]]<|/det|>
+3. Which phenotypes do monocytes and macrophages acquire at 6 h and 3 days after intrapalantar IL-6 and after PGE2? Are they more likely to release pro-nociceptive chemicals? (Figure 6).  
+
+<|ref|>text<|/ref|><|det|>[[115, 399, 789, 435]]<|/det|>
+4. How do the authors explain IL-6 and PGE2 pro-nociceptive effects at 1 h after injections in NAAACD11b-/- in figure 3?  
+
+<|ref|>text<|/ref|><|det|>[[116, 474, 392, 491]]<|/det|>
+Reviewer #3 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 503, 876, 594]]<|/det|>
+In this manuscript, the authors aim to identify molecular mechanism of acute to chronic pain transition. They concluded a critical role of NAAA on monocyte- derived cells and PPAR- a receptors in the event by using a mouse model of hyperalgesic priming. The study is carefully designed, with various experimental approaches. While a large amount of data supported most of their conclusion, I have several major concerns on the limitation of the study.  
+
+<|ref|>text<|/ref|><|det|>[[115, 632, 877, 761]]<|/det|>
+- Model of hyperalgesic priming: I greatly appreciate the model provides us an evidence that prior priming nociceptors could lead increased sensitivity to subsequent stimulation. The authors in this study demonstrated potential underlying mechanism mediated by monocyte-derived cells-associated NAAA. However, this is a very restricted model, only limited to inflammatory triggers, with specific agents and specific doses. I am sure not if the agent or the dose changed, whether the 72hour-interval is still valid, not saying if this is valid with a non-inflammatory trigger. It may mislead knowledge users by generalize the findings.  
+
+<|ref|>text<|/ref|><|det|>[[115, 771, 875, 825]]<|/det|>
+- Pain behavioral testing methods: Only heat sensitivity was assessed in the study. Whether such priming is also effective in mechanical and cold sensitivity is unknown. Whether NAAA mediates other pain modalities following the priming was not assessed either in the study.  
+
+<|ref|>text<|/ref|><|det|>[[115, 836, 879, 890]]<|/det|>
+- Skin macrophages: Two experimental approaches to deplete monocytes used in the study are indeed not ideal. Both are more effective and suitable in depleting macrophages/microglia than monocytes. The authors repeatedly indicated that observed effects are derived from monocytes/macrophages. They
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 90, 880, 162]]<|/det|>
+excluded CNS microglia effect, but they didn't touch anything on skin macrophages, which is essential. If we believe there is a crucial peripheral contribution of NAAA in inflammatory mediator induced priming, detailed changes of macrophages in affected paw are indispensable, as they are direct players. Such data may also help to understand why it works not before or not after 72hr interval.  
+
+<|ref|>text<|/ref|><|det|>[[115, 173, 872, 227]]<|/det|>
+- Clinical implications: It is difficult for me to imagine in what clinical setting where the findings from the current study can apply. What does this 72 hr interval, not before or not after, mean for patients who come to see a physician for an acute pain?  
+
+<|ref|>text<|/ref|><|det|>[[115, 266, 881, 338]]<|/det|>
+Thus, in general, I recognize the value and the quality of the study in identifying molecular mechanism of hypersensitivity priming in a specific setting, I feel the conclusion, including the title of the study was overstated. Chronic pain is complex, I believe we need to develop precision medicine for each specific type of chronic pain.  
+
+<|ref|>text<|/ref|><|det|>[[117, 378, 392, 394]]<|/det|>
+Reviewer #4 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[114, 406, 881, 570]]<|/det|>
+The manuscript "NAAA- regulated lipid signaling in monocyte- derived cells controls the induction of hyperalgesic priming", by Fotio et al, provides additional mechanistic information regarding a model of the transition from acute to chronic pain, hyperalgesic priming (HP). This is an interesting study, with the experiments appropriately conducted, clearly exemplifying the involvement of the immune system in the development of persistent nociceptive sensitization. The conclusions certainly add to the characterization of such phenomenon, previously shown to play a role in the persistence of pain observed in some clinical conditions. Although I do not see a reason preventing its publication, I would appreciate if the authors could clarify or include some minor comments in the final version of the manuscript:  
+
+<|ref|>text<|/ref|><|det|>[[114, 581, 868, 709]]<|/det|>
+- Experimentally, HP is defined by the potentiation (prolongation) of the mechanical hyperalgesia induced by PGE2 (>4h) in the paw when compared to a non-primed paw (<2h), indicating the hyper-responsive state triggered by the inflammation in the primary neuron. In the current study the authors evaluate the effect of PGE2 on the thermal nociceptive threshold in primed paws. What is the time course of the response produced by PGE2 on the thermal sensitivity in a normal, non-primed paw? Including this information would help the reader to visualize the increased sensitivity of the nociceptor produced by the priming stimulus, for comparison.  
+
+<|ref|>text<|/ref|><|det|>[[114, 719, 876, 846]]<|/det|>
+- Induction of HP by carrageenan is, in rats, sexually dimorphic, modulated by estrogen at the level of PKCepsilon. In mice, HP is also sexually dimorphic. However, instead of being an "all or nothing" phenomenon as in rats, this difference in mice seems to be more complex. Can the authors include a comment about these differences in the Discussion? Since the development of HP in mice of both sexes is suggested to occur at the level of NAAA, the upstream differential mechanism could be mentioned (or speculated), specifically considering that NAAA is important only during the incubation phase, when protein translation is happening.  
+
+<|ref|>text<|/ref|><|det|>[[113, 857, 852, 893]]<|/det|>
+- Why did the authors use the tail flick method to determine nociceptive thresholds if all the relevant behavior experiments were performed using the Hargreaves plantar test in the paw?
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 90, 676, 107]]<|/det|>
+- Were the compounds injected in the paw intradermally or subcutaneously?  
+
+<|ref|>text<|/ref|><|det|>[[115, 118, 872, 172]]<|/det|>
+The information provided by this study contribute to the knowledge about the mechanisms involved in the transition from acute to chronic pain, which is of utmost importance. I look forward to receiving the authors' response and comments added to the revised version of this article.
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[327, 89, 670, 107]]<|/det|>
+## RESPONSE TO REVIEWER COMMENTS  
+
+<|ref|>text<|/ref|><|det|>[[115, 121, 882, 170]]<|/det|>
+We are grateful to the Reviewers for their constructive and stimulating comments. We addressed them with several new experiments and substantive text edits, highlighted in red in this resubmission.  
+
+<|ref|>sub_title<|/ref|><|det|>[[115, 209, 219, 225]]<|/det|>
+## Reviewer #1  
+
+<|ref|>text<|/ref|><|det|>[[115, 240, 882, 466]]<|/det|>
+N- acylethanolamine- hydrolyzing acid amidase (NAAA) is a lysosomal enzyme abundant in monocytes and macrophages. NAAA hydrolyzes N- acylethanolamines, including palmitoylethanolamide (PEA) and oleoylethanolamide (OEA). Over the years, the authors have studied the anti- inflammatory and analgesic actions of N- acylethanolamines in vivo and have developed specific NAAA inhibitors which exhibit anti- inflammatory and analgesic effects. In the present study the authors extended their interests to the involvement of NAAA in hyperalgesic priming. They used specific NAAA inhibitors and plural kinds of NAAA- deficient mice to provide the evidence that NAAA- regulated signaling at PPAR- alpha directs monocytes and macrophages to initiate hyperalgesic priming in mice exposed to an inflammatory stimulus. This is the first report clearly showing that NAAA- regulated signaling is involved in hyperalgesic priming. Thus, the work is of significance to the field and rich in originality. The work sufficiently supports the conclusions. There are no flaws in the data analysis, interpretation and conclusions. The methodology is also sound. I think that the work meets the expected standards in our field. I just have some minor comments.  
+
+<|ref|>text<|/ref|><|det|>[[173, 480, 608, 497]]<|/det|>
+We are grateful for this positive evaluation of our work.  
+
+<|ref|>text<|/ref|><|det|>[[115, 511, 792, 529]]<|/det|>
+1. In Fig. 4E, the symbols for significance should be placed near the line of CD11b-/-.  
+
+<|ref|>text<|/ref|><|det|>[[115, 544, 610, 561]]<|/det|>
+2. In Fig. 4F1, the asterisks should be explained in the legend.  
+
+<|ref|>text<|/ref|><|det|>[[115, 576, 476, 593]]<|/det|>
+3. In Fig. 4G1, the symbols may be asterisks.  
+
+<|ref|>text<|/ref|><|det|>[[174, 608, 222, 624]]<|/det|>
+Done.  
+
+<|ref|>sub_title<|/ref|><|det|>[[115, 656, 221, 673]]<|/det|>
+## Reviewer #2  
+
+<|ref|>text<|/ref|><|det|>[[116, 688, 882, 800]]<|/det|>
+This study shows that the induction of hyperalgesic priming by intraplantar IL- 6 to intraplantar PGE2 involves monocytes and macrophages. These immune cells are the source for hydrolase NAAA that acts on two lipids (PEA and OEA) which are agonists for intracellular PPAR- alpha signalling. The interruption of this pathway in monocytes and macrophages by a peripheralized NAAA inhibitor blocks the induction of IL- 6- mediated hyperalgesic priming in male and female mice. In addition, HP induction is impaired in global NAAA knockout and following silencing of NAAA in CD11b cells and by clodronate liposome and CSF1 receptor antagonist treatments.  
+
+<|ref|>text<|/ref|><|det|>[[116, 815, 882, 896]]<|/det|>
+These data suggest that monocyte/macrophages are critical for the induction of hyperalgesic priming and in such cells lipid signalling regulated by NAAA plays a mechanistic role. This is an interesting possibility for a novel neuro- immune interaction- mediated mechanisms in the induction of persistent pain. However, the current behavioural data in transgenic mice should be paired with characterization of monocytes/macrophages in the injected paw before and after
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 88, 883, 170]]<|/det|>
+intraplantar injections of IL- 6 and PGE2. More specifically some of the following points should be addressed: 1. Do hind paw macrophages express and regulate expression of IL- 6 receptor? What levels of NAAA, PEA, OEA and palmitic acid do they contain before, 6 hours and 3 days after IL- 6 injections? 2. What is the level of NAAA, PEA, OEA and PA in macrophages isolated from the hind paw of NAAACD11b-/- and PparaCD11b-/- mice?  
+
+<|ref|>text<|/ref|><|det|>[[165, 200, 883, 345]]<|/det|>
+We thank the Reviewer for this important comment, which was echoed by a similar one made by Reviewer 3. Prompted by these suggestions, we conducted a new experiment to assess whether resident macrophages contribute to the initiation of hyperalgesic priming. We removed macrophages from the hind paws of mice with two intraplantar injections of clodronate liposomes (PBS liposomes were injected as control) and examined the effects of IL- 6 administration in the same paws. The results, which are reported in new Figure S9, show that clodronate treatment substantially reduced the number of \(\mathrm{CD68^{+}}\) macrophages in paw tissue (Fig. S9B) but did not prevent IL- 6- induced hyperalgesic priming (new Fig. S9E).  
+
+<|ref|>image<|/ref|><|det|>[[333, 350, 723, 500]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[180, 510, 870, 567]]<|/det|>
+From Fig. S9. (B) Effect of PBS- containing liposomes (left) and clodronate- containing liposomes (right) on \(\mathrm{CD68 + }\) cells (resident macrophages) in mouse paw skin. (C) Effect of PBS- or clodronate- containing liposomes on the induction of hyperalgesic priming (response to PGE2 six days after IL- 6). See Figure S9 for additional details.  
+
+<|ref|>text<|/ref|><|det|>[[166, 589, 881, 623]]<|/det|>
+By contrast, clodronate delayed the acute response to IL- 6, suggesting that local macrophages contribute in part to such response.  
+
+<|ref|>image<|/ref|><|det|>[[413, 640, 627, 772]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[180, 782, 866, 812]]<|/det|>
+From Fig. S9. (E) Effect of PBS- or clodronate- containing liposomes on the acute nociceptive response to IL- 6. # \(P < 0.05\) compared to PBS ( \(n = 8\) ). See Figure S9 for details.  
+
+<|ref|>text<|/ref|><|det|>[[166, 835, 882, 901]]<|/det|>
+Along with the data shown in Figure 5 (effects of systemic clodronate and PLX- 5622) these findings indicate that blood- borne monocytes, rather than resident macrophages, are required for priming initiation. We modified our hypothetical model (and the title of the study) to reflect these new results (please see new Fig. 7).
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 105, 883, 139]]<|/det|>
+Which phenotypes do monocytes and macrophages acquire at 6 h and 3 days after intraplantar IL- 6 and after PGE2? Are they more likely to release pro- nociceptive chemicals? (Figure 6).  
+
+<|ref|>text<|/ref|><|det|>[[165, 153, 883, 281]]<|/det|>
+Based on our results showing that monocytes, not resident macrophages, are involved in priming, we addressed this question by examining the molecular phenotype of circulating immune cells by high- resolution mass cytometry by time of flight (CyTOF). Mice received intraplantar IL- 6 injections and, 72 hours later (i.e., at the tail end of the incubation phase of priming), cardiac blood was collected for CyTOF analyses. The results, which are illustrated in new Figure 6, indicate that the incubation phase of hyperalgesic priming coincides with the emergence of one or more subpopulation(s) of activated circulating monocytes. For example:  
+
+<|ref|>image<|/ref|><|det|>[[384, 297, 603, 550]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[212, 574, 901, 617]]<|/det|>
+From Fig. 6. (B) Optimized Stochastic Neighbor Embedding (opt- SNE) plots depicting density heatmaps of cells expressing the markers of monocyte activation CCR2 (top), CD43 (middle), and CX3CR1 (bottom) in vehicle- (left) and IL- 6- treated (right) mice. See new Figure 6 for details.  
+
+<|ref|>text<|/ref|><|det|>[[115, 658, 884, 691]]<|/det|>
+4. How do the authors explain IL-6 and PGE2 pro-nociceptive effects at 1 h after injections in NAAACD11b-/- in figure 3?  
+
+<|ref|>text<|/ref|><|det|>[[163, 705, 883, 737]]<|/det|>
+As pointed out by the Reviewer, Figure 3E shows that, one h after injection, PGE2 exerts a nociceptive effect in both control Naaa f/l and Naaa CD11b-/- mice:  
+
+<|ref|>image<|/ref|><|det|>[[425, 752, 549, 870]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[127, 872, 870, 901]]<|/det|>
+From Fig. 3. (E) Acute nociceptive effects of PGE2 in Naaa CD11b-/- mice (green symbols) and control Naaa f/l mice (magenta symbols) #P < 0.001 (n = 8). See Figure 3 for details.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[165, 105, 882, 155]]<|/det|>
+This effect is expected because the 1h time point measures the acute nociceptive response to \(\mathsf{PGE}_2\) , which is generally not affected by pharmacological or genetic interventions targeting NAAA. See, for example:  
+
+<|ref|>image<|/ref|><|det|>[[123, 171, 880, 305]]<|/det|>
+<|ref|>image_caption<|/ref|><|det|>[[125, 314, 864, 358]]<|/det|>
+<center>From Fig. 2. (A-E) Acute nociceptive effects of \(\mathsf{PGE}_2\) in IL6-treated mice. Note the persistence of the effects at the 1-h time point even in mice treated with NAAA inhibitor ARN19702 (please see Fig. 2 for details). #P < 0.05, #P < 0.01, and #P < 0.001 compared to veh/veh (n = 10). See Figure S9 for details. </center>  
+
+<|ref|>text<|/ref|><|det|>[[165, 379, 882, 428]]<|/det|>
+One exception is represented by global Naaa ko mice, which exhibited a weaker response to \(\mathsf{PGE}_2\) at the 1- h time point (Fig. 3). This could be due to NAAA deletion in as- yet- unidentified cells that are not readily accessible to the NAAA inhibitors used in our study.  
+
+<|ref|>sub_title<|/ref|><|det|>[[116, 443, 221, 460]]<|/det|>
+## Reviewer #3  
+
+<|ref|>text<|/ref|><|det|>[[115, 474, 882, 556]]<|/det|>
+In this manuscript, the authors aim to identify molecular mechanism of acute to chronic pain transition. They concluded a critical role of NAAA on monocyte- derived cells and PPAR- a receptors in the event by using a mouse model of hyperalgesic priming. The study is carefully designed, with various experimental approaches. While a large amount of data supported most of their conclusion, I have several major concerns on the limitation of the study.  
+
+<|ref|>text<|/ref|><|det|>[[115, 570, 882, 684]]<|/det|>
+Model of hyperalgesic priming: I greatly appreciate the model provides us an evidence that prior priming nociceptors could lead increased sensitivity to subsequent stimulation. The authors in this study demonstrated potential underlying mechanism mediated by monocyte- derived cells- associated NAAA. However, this is a very restricted model, only limited to inflammatory triggers, with specific agents and specific doses. I am sure not if the agent or the dose changed, whether the 72hour- interval is still valid, not saying if this is valid with a non- inflammatory trigger. It may mislead knowledge users by generalize the findings.  
+
+<|ref|>text<|/ref|><|det|>[[157, 698, 882, 884]]<|/det|>
+Hyperalgesic priming can be induced by a variety of stimuli, including proinflammatory cytokines (e.g., IL- 6, TNF- \(\alpha\) ), growth factors (e.g., NGF, CSF- 1), inflammatory triggers (e.g., carrageenan), and tissue damage (e.g., surgical paw incision). Its clinical relevance has been convincingly argued by various investigators (Levine, Price, etc.) who pointed out that "... the experimental framework of the hyperalgesic priming model provides important insight into clinical chronic pain because it captures the recurrent nature of some of the most common pathological pain conditions." (Kandasamy R, Price TJ. The pharmacology of nociceptor priming. Handb Exp Pharmacol. 2015; 227:15- 37. PMID: 25846612) Our previous submission described the results of experiments with three distinct proinflammatory stimuli – IL- 6, TNF- \(\alpha\) , and carrageenan. Nevertheless, to address the spirit of the Reviewer's concern, we have included new data using a fourth stimulus, paw incision, which has a strong nociceptive component. The results, reported in new Figure S7, show
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[156, 88, 882, 123]]<|/det|>
+that administration of a NAAA inhibitor in the first 72h after the incision fully prevents priming initiation, lending further support to the general validity of our findings.  
+
+<|ref|>text<|/ref|><|det|>[[117, 136, 882, 186]]<|/det|>
+Pain behavioral testing methods: Only heat sensitivity was assessed in the study. Whether such priming is also effective in mechanical and cold sensitivity is unknown. Whether NAAA mediates other pain modalities following the priming was not assessed either in the study.  
+
+<|ref|>text<|/ref|><|det|>[[150, 201, 882, 234]]<|/det|>
+We assessed both heat and mechanical hypersensitivity, with similar results. Representative sets of mechanical hypersensitivity data are included in new Figure S2 and S4.  
+
+<|ref|>text<|/ref|><|det|>[[117, 249, 882, 298]]<|/det|>
+Skin macrophages: Two experimental approaches to deplete monocytes used in the study are indeed not ideal. Both are more effective and suitable in depleting macrophages/microglia than monocytes.  
+
+<|ref|>text<|/ref|><|det|>[[152, 313, 882, 361]]<|/det|>
+We direct the Reviewer's attention to the results reported in Figure 5, which show that both approaches (clodronate and PLX- 5622) substantially reduced monocyte numbers in the bloodstream.  
+
+<|ref|>text<|/ref|><|det|>[[116, 376, 883, 473]]<|/det|>
+The authors repeatedly indicated that observed effects are derived from monocytes/macrophages. They excluded CNS microglia effect, but they didn't touch anything on skin macrophages, which is essential. If we believe there is a crucial peripheral contribution of NAAA in inflammatory mediator induced priming, detailed changes of macrophages in affected paw are indispensable, as they are direct players. Such data may also help to understand why it works not before or not after 72hr interval.  
+
+<|ref|>text<|/ref|><|det|>[[152, 489, 882, 521]]<|/det|>
+We thank the Reviewer for this suggestion, which we have addressed with the new experiment described in our response to Reviewer 2. Please see above.  
+
+<|ref|>text<|/ref|><|det|>[[116, 536, 882, 585]]<|/det|>
+Clinical implications: It is difficult for me to imagine in what clinical setting where the findings from the current study can apply. What does this 72 hr interval, not before or not after, mean for patients who come to see a physician for an acute pain?  
+
+<|ref|>text<|/ref|><|det|>[[152, 600, 882, 681]]<|/det|>
+This basic science study is not meant to have immediate repercussions on clinical practice. It does have, however, potential clinical significance in that it suggests that postsurgical NAAA inhibition (but not postsurgical treatment with standard analgesics) may prevent the transition to pain chronicity, a significant problem for patients who undergo thoracotomy, knee arthroplasty, mastectomy, and other invasive surgeries.  
+
+<|ref|>text<|/ref|><|det|>[[116, 696, 882, 760]]<|/det|>
+Thus, in general, I recognize the value and the quality of the study in identifying molecular mechanism of hypersensitivity priming in a specific setting, I feel the conclusion, including the title of the study was overstated. Chronic pain is complex, I believe we need to develop precision medicine for each specific type of chronic pain.  
+
+<|ref|>text<|/ref|><|det|>[[150, 775, 882, 808]]<|/det|>
+Our views about the complexity of chronic pain are very similar to the Reviewer's, as affirmed in the Discussion section of the manuscript:  
+
+<|ref|>text<|/ref|><|det|>[[152, 823, 882, 888]]<|/det|>
+"Chronic pain states are widely heterogeneous in causes, symptoms, impact on function, and temporal development<sup>3</sup>. This diversity justifies skepticism toward a simplistic view of the progression to pain chronicity as transformation of one mechanistic type of pain (e.g., acute pain associated with injury) into another (e.g., neuropathic pain)<sup>99</sup>."
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[151, 89, 883, 218]]<|/det|>
+Thus, we are surprised by the Reviewer's statement that "the conclusion, including the title of the study was overstated." The title of our manuscript (NAAA- regulated lipid signaling in monocytes controls the induction of hyperalgesic priming) accurately reflects the findings presented. Our conclusion is equally cautious and, we believe, fully warranted by the data: "Clinical studies should determine whether algostatic agents might offer a strategy to prevent chronic pain after invasive surgery and other kinds of physical trauma." However, when interpretation could have been ambiguous, we replaced 'chronic pain' with 'persistent pathological nociception' (see, for example, Abstract, line 38).  
+
+<|ref|>sub_title<|/ref|><|det|>[[116, 256, 220, 273]]<|/det|>
+## Reviewer #4  
+
+<|ref|>text<|/ref|><|det|>[[116, 288, 883, 401]]<|/det|>
+The manuscript "NAAA- regulated lipid signaling in monocyte- derived cells controls the induction of hyperalgesic priming", by Fotio et al, provides additional mechanical information regarding a model of the transition from acute to chronic pain, hyperalgesic priming (HP). This is an interesting study, with the experiments appropriately conducted, clearly exemplifying the involvement of the immune system in the development of persistent nociceptive sensitization. The conclusions certainly add to the characterization of such phenomenon, previously shown to play a role in the persistence of pain observed in some clinical conditions.  
+
+<|ref|>text<|/ref|><|det|>[[160, 416, 464, 433]]<|/det|>
+We appreciate the positive comments.  
+
+<|ref|>text<|/ref|><|det|>[[116, 447, 883, 481]]<|/det|>
+Although I do not see a reason preventing its publication, I would appreciate if the authors could clarify or include some minor comments in the final version of the manuscript:  
+
+<|ref|>text<|/ref|><|det|>[[116, 495, 883, 609]]<|/det|>
+Experimentally, HP is defined by the potentiation (prolongation) of the mechanical hyperalgesia induced by PGE2 (>4h) in the paw when compared to a non- primed paw (<2h), indicating the hyper- responsive state triggered by the inflammation in the primary neuron. In the current study the authors evaluate the effect of PGE2 on the thermal nociceptive threshold in primed paws. What is the time course of the response produced by PGE2 on the thermal sensitivity in a normal, non- primed paw? Including this information would help the reader to visualize the increased sensitivity of the nociceptor produced by the priming stimulus, for comparison.  
+
+<|ref|>text<|/ref|><|det|>[[156, 623, 882, 657]]<|/det|>
+We have added a new supplemental figure (Fig. S2) that contains the requested information.  
+
+<|ref|>text<|/ref|><|det|>[[116, 671, 883, 785]]<|/det|>
+Induction of HP by carrageenan is, in rats, sexually dimorphic, modulated by estrogen at the level of PKCepsilon. In mice, HP is also sexually dimorphic. However, instead of being an "all or nothing" phenomenon as in rats, this difference in mice seems to be more complex. Can the authors include a comment about these differences in the Discussion? Since the development of HP in mice of both sexes is suggested to occur at the level of NAAA, the upstream differential mechanism could be mentioned (or speculated), specifically considering that NAAA is important only during the incubation phase, when protein translation is happening.  
+
+<|ref|>text<|/ref|><|det|>[[160, 800, 882, 865]]<|/det|>
+We do not know how inflammatory challenges trigger NAAA activation and whether this process might be sexually dimorphic. But we appreciate the point raised by the Reviewer and addressed its spirit by adding the following comment to the Discussion section of the manuscript (p. 17):  
+
+<|ref|>text<|/ref|><|det|>[[157, 871, 882, 904]]<|/det|>
+"In this context, an important question that remains to be answered pertains to the mechanisms through which inflammatory stimuli lead to the suppression of NAPE- PLD
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[160, 89, 882, 138]]<|/det|>
+transcription and the enhancement of NAAA activity. Elucidating such mechanisms might shed light on aspects of HP – such as the existence of sexual dimorphisms in rats<sup>88- 90</sup> and mice<sup>91</sup> – which are presently unclear."  
+
+<|ref|>text<|/ref|><|det|>[[118, 160, 883, 194]]<|/det|>
+Why did the authors use the tail flick method to determine nociceptive thresholds if all the relevant behavior experiments were performed using the Hargreaves plantar test in the paw?  
+
+<|ref|>text<|/ref|><|det|>[[159, 208, 882, 273]]<|/det|>
+We routinely use the tail- flick method to identify potential alterations in baseline nociceptive threshold in newly developed mouse lines such as those introduced in the present study. Here, we combined tail- flick (Fig. S9) and Hargreaves plantar tests (Fig. 3D and Fig. 4G1) to obtain a more complete view of the lines' nociceptive phenotype.  
+
+<|ref|>text<|/ref|><|det|>[[118, 288, 712, 305]]<|/det|>
+Were the compounds injected in the paw intradermally or subcutaneously?  
+
+<|ref|>text<|/ref|><|det|>[[159, 320, 882, 369]]<|/det|>
+Compounds were injected subcutaneously, between skin and muscular fascia/tendon, as customary in this model. We clarified this point in the Materials and Methods section of the manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[117, 383, 883, 431]]<|/det|>
+The information provided by this study contribute to the knowledge about the mechanisms involved in the transition from acute to chronic pain, which is of utmost importance. I look forward to receiving the authors' response and comments added to the revised version of this article.  
+
+<|ref|>text<|/ref|><|det|>[[159, 447, 249, 464]]<|/det|>
+Thank you.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 145, 393, 163]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 174, 697, 192]]<|/det|>
+The authors appropriately revised the manuscript in response to my comments.  
+
+<|ref|>text<|/ref|><|det|>[[115, 230, 393, 248]]<|/det|>
+Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 259, 416, 277]]<|/det|>
+My points have been partially addressed.  
+
+<|ref|>text<|/ref|><|det|>[[115, 287, 872, 380]]<|/det|>
+Specifically, whilst I suggested to characterise monocytes/macrophages at the site of IL- 6 injection (hind paw) the authors have looked at blood monocytes. They provide evidence that intraplantar injection of IL- 6 is associated with more CCR2+ monocytes circulating in blood. However, classical monocytes will have to infiltrate tissue to come in the vicinity of primary afferent terminals and promote nociceptive signalling.  
+
+<|ref|>text<|/ref|><|det|>[[115, 389, 864, 444]]<|/det|>
+Whether CCR2+ monocytes infiltrate the site of injection remains to be established and Figure 7 schematic should indicate that this study does not provide direct evidence that "monocytes migrate to target tissue".  
+
+<|ref|>text<|/ref|><|det|>[[115, 454, 842, 472]]<|/det|>
+New Figure 6: How do the authors explain presence of F4/80 positive cells (macrophages) in blood?  
+
+<|ref|>text<|/ref|><|det|>[[115, 482, 854, 537]]<|/det|>
+Discussion: It is possible that CCL2 is upregulated in endothelial cells, and sensory neurons and accumulates in primary afferent terminals. However, the authors provide no direct evidence that this happens in their experimental conditions.  
+
+<|ref|>text<|/ref|><|det|>[[115, 547, 875, 602]]<|/det|>
+Other points. Ref 12 line 57: This is an early observation and more recent evidence indicates no infiltration of monocytes in the spinal cord following peripheral nerve injury. The authors could rephrase their statement.  
+
+<|ref|>text<|/ref|><|det|>[[115, 612, 438, 630]]<|/det|>
+Line 167: replace "deleted" with "depleted".  
+
+<|ref|>text<|/ref|><|det|>[[115, 668, 393, 686]]<|/det|>
+Reviewer #3 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 697, 696, 715]]<|/det|>
+The authors addressed the reviewer's concerns. I don't have further comments.  
+
+<|ref|>text<|/ref|><|det|>[[115, 754, 393, 772]]<|/det|>
+Reviewer #4 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[112, 783, 813, 801]]<|/det|>
+I am satisfied with the revision provided by the authors. I recommend this study for publication.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[85, 88, 934, 187]]<|/det|>
+Specifically, whilst I suggested to characterise monocytes/macrophages at the site of IL- 6 injection (hind paw) the authors have looked at blood monocytes. They provide evidence that intraplantar injection of IL- 6 is associated with more CCR2+ monocytes circulating in blood. However, classical monocytes will have to infiltrate tissue to come in the vicinity of primary afferent terminals and promote nociceptive signalling. Whether CCR2+ monocytes infiltrate the site of injection remains to be established and Figure 7 schematic should indicate that this study does not provide direct evidence that "monocytes migrate to target tissue".  
+
+<|ref|>text<|/ref|><|det|>[[85, 201, 933, 235]]<|/det|>
+We clarified that Figure 7 provides a hypothetical model which summarizes the data presented in our study and suggests a roadmap for future experiments. Please see Figure 7 legend.  
+
+<|ref|>text<|/ref|><|det|>[[88, 248, 930, 266]]<|/det|>
+New Figure 6: How do the authors explain presence of F4/80 positive cells (macrophages) in blood?  
+
+<|ref|>text<|/ref|><|det|>[[85, 280, 930, 314]]<|/det|>
+Though unlikely, it is possible that the antibody we used cross- reacted with a different monocyte antigen. Owing to this uncertainty, we removed F4/80 data from Figure 6C.  
+
+<|ref|>text<|/ref|><|det|>[[85, 328, 933, 378]]<|/det|>
+Discussion: It is possible that CCL2 is upregulated in endothelial cells, and sensory neurons and accumulates in primary afferent terminals. However, the authors provide no direct evidence that this happens in their experimental conditions.  
+
+<|ref|>text<|/ref|><|det|>[[85, 392, 930, 425]]<|/det|>
+There is substantial evidence that peripheral nerve injury upregulates CCL2 in DRG neurons, but future experiments will have to address this question as it pertains to the model used here.  
+
+<|ref|>text<|/ref|><|det|>[[85, 440, 930, 473]]<|/det|>
+Other points. Ref 12 line 57: This is an early observation and more recent evidence indicates no infiltration of monocytes in the spinal cord following peripheral nerve injury. The authors could rephrase their statement.  
+
+<|ref|>text<|/ref|><|det|>[[85, 488, 933, 568]]<|/det|>
+Evidence indicates that spinal nerve resection does not stimulate monocytes migration toward the spinal cord (PMID 27373153, presumably the study the reviewer had in mind). The hyperalgesic priming model we used, however, is quite different and, possibly, mechanistically closer to the one used in Ref 12 (partial sciatic nerve ligation). Indeed, both hyperalgesic priming and sciatic nerve ligation exhibit a rather strong inflammatory component that is not observed after spinal nerve resection.  
+
+<|ref|>text<|/ref|><|det|>[[86, 583, 429, 600]]<|/det|>
+Line 167: replace "deleted" with "depleted".  
+
+<|ref|>text<|/ref|><|det|>[[86, 616, 135, 631]]<|/det|>
+Done.
+
+<--- Page Split --->

peer_reviews/12438c2ece72c0a7ec2a093dc88eaba15e637ea325a5525049bbab0bfceb822d/supplementary_0_Peer Review File/images_list.json

@@ -0,0 +1,41 @@
+[
+  {
+    "type": "image",
+    "img_path": "images/Figure_4.jpg",
+    "caption": "Figure 4 | Total Cr(VI) in wind-dispersible soil and ash particles. a. \\(\\mu\\) -XRF image (pixel resolution: \\(1 \\mu m\\) ) showing the relative intensity of Cr(VI) (green; estimated as the intensity ratio at 5993 and 6010 eV) and total Cr (blue; measured at 6010 eV) within the \\(< 53 - \\mu m\\) size fraction of Cr-bearing soil-ash particulates from a serpentine chaparral that experienced high fire severity (A7). b. Normalized \\(\\mu\\) -XANES spectra (Cr K-edge) from numbered locations on Cr-bearing particles in a. Dashed lines indicate energies characteristic of Cr(VI) (5993 eV), Cr(III) (6003 eV), and total Cr (6010 eV), at which \\(\\mu\\) -XRF images were also collected.",
+    "footnote": [],
+    "bbox": [],
+    "page_idx": 0
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_0.jpg",
+    "caption": "Figure S9 | Total Cr(VI) in wind-dispersible soil and ash particles. a. \\(\\mu\\) -XRF image (pixel resolution: \\(1\\mu \\mathrm{m}\\) ) showing the relative intensity of \\(\\mathrm{Cr(VI)}\\) (green; estimated as the intensity ratio at 5993 and \\(6010\\mathrm{eV}\\) ) and total Cr (blue; measured at \\(6010\\mathrm{eV}\\) ) within the \\(< 53 - \\mu \\mathrm{m}\\) size fraction of Cr-bearing soil-ash particulates from a serpentine chaparral that experienced high fire severity (A7). b. Normalized \\(\\mu\\) -XANES spectra (Cr K-edge) from numbered locations on Cr-bearing particles in a. Dashed lines indicate energies characteristic of \\(\\mathrm{Cr(VI)}\\) (5993 eV), \\(\\mathrm{Cr(III)}\\) (6003 eV), and total Cr (6010 eV).",
+    "footnote": [],
+    "bbox": [
+      [
+        128,
+        182,
+        777,
+        444
+      ]
+    ],
+    "page_idx": 9
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_1.jpg",
+    "caption": "Figure S7: Legend : ... from a serpentine chaparral soil...",
+    "footnote": [],
+    "bbox": [],
+    "page_idx": 10
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_4.jpg",
+    "caption": "Figure 4: Change A and B in Figure plot panels to a and b reflected in the figure caption",
+    "footnote": [],
+    "bbox": [],
+    "page_idx": 10
+  }
+]

peer_reviews/12438c2ece72c0a7ec2a093dc88eaba15e637ea325a5525049bbab0bfceb822d/supplementary_0_Peer Review File/supplementary_0_Peer Review File.mmd

@@ -0,0 +1,389 @@
+
+# nature portfolio  
+
+# Peer Review File  
+
+# Metal Toxin Threat in Wildland Fires Determined by Geology and Fire Severity  
+
+![](images/Figure_4.jpg)
+  
+
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+
+# REVIEWER COMMENTS  
+
+Reviewer #1 (Remarks to the Author):  
+
+# Review of manuscript 395109 submitted to Nature Communication  
+
+# Metal toxin threat in wildland fires determined by geology and fire severity  
+
+Alandra Marie Lopez, Juan Lezama Pacheco and Scott Fendorf  
+
+This manuscript presents the results of a study aimed at emphasizing the health threat potentially arising from chromium transformation to its most harmful \(\mathrm{Cr(VI)}\) form in soils and ashes upon wildfires. To reach their objective, the authors analyzed 38 cores drilled in more or less severely burned soils at four Preserves that experienced large wildfires in the North Coast Range of California (USA). The results obtained confirm the catalytic effect of high temperatures generated by wildfires on chromium oxidation in soil, and they point to control of soil geology and fire severity on this effect. They also interestingly indicate that reactive \(\mathrm{Cr(VI)}\) can reach dangerous levels in wind- dispersible particulates found in the surficial layers of the soil at ultramafic settings, and that this harmful form of chromium can persist in the soil/ash system up to one year after the wildfire if rainfalls are not significant. These results yield to the conclusion that more work is required to further evaluate this potential risk around the world.  
+
+I found this paper very pleasant to read. The subject is well introduced, the sites and samples are well described and fit with the objective of the study, the results are well presented and they support the discussion that gives proper attention to the existing literature on the topic and yield to the novel conclusions that geology and fire intensity are important drivers of harmful hexavalent chromium in soils and wind- dispersible soil/ash surface particles, and that the associated health risk can persist up to one year after fire.  
+
+Although the scale of the study (North Coast Range of California, USA) could at first be considered too short to extend its findings at the global scale, I agree with the authors that the number and diversity of sites studied can be considered enough to address this issue. I also agree with the concluding remark on the importance of further evaluating the potential risk of wildfire- induced harmful hexavalent chromium in wind- dispersible soil/ash particles at the global scale. Finally, I agree with the authors that the results provided in this study deserve to be shared with a large audience, ranging from scientists working on the topic to policy makers, public administrators and nature managers.  
+
+For all these reasons, I consider that this manuscript deserves to be published in Nature Communications. I have listed below few issues that I would be interested to see addressed, although only few of them are mandatory for publication.  
+
+## Figure 1  
+
+I agree that a large fraction of ultramafic/mafic areas are concerned by wildfires at the global scale. Figure 1 shows that these areas are mainly located in the tropical region where most soils are deeply weathered (Utilisols and Oxisols, according to the USDA classification). However, the Fe contents reported in Table S6 suggest that the soils studied in this work do not correspond to these soil types. This raises the question of the actual representativeness of the results regarding tropical areas. I would
+
+<--- Page Split --->
+
+recommend that the authors comment on that point, and maybe further consider it in their concluding remarks.  
+
+## Figure 3  
+
+Maybe, the authors could add the results of comparative statistics tests on barplots and boxplots (either directly report the p- values or show corresponding labels as \*\*\* or \*\*\*)?  
+
+Do they have any hypothesis to explain the much higher concentration in reactive C(VI) measured in the A7 surface soil- ash sample from the serpentine chaparral landscape ? The data provided in Table S6 indicate about twice more total Cr in this sample compared to sample A6 (for instance), but the concentration in reactive Cr(VI) is more than 3 times higher in the bulk fraction and more than 25 three times higher in the \(< 53\mu m\) fraction. Are there any mineralogical differences with the other moderatelyhighly burned serpentine soil- ash samples (A3- A6) that could help to explain that ?  
+
+## Figure 4  
+
+I wonder if this figure should be maintained in the main text. First, I am not convince that it really supports the assumption that 'total Cr(VI) was most abundant in wind- dispersible soils and ash particulates after high fire severity conditions compared to a low severity sample'. Second, I do not understand why the two figures are plotted on different sizes (A is larger than B).  
+
+It could maybe be replaced by a figure similar to Figure S7 (maybe use only panels A and B or panels C and D from Figure S7 and add two similar panels from a low fire severity burn sample from a serpentine chaparral soil)?  
+
+Whatever, if Figure 4 is maintained in the main text, I would recommend that the authors add some histograms with the estimated number of Cr(VI) and total Cr particles on the two figures, in order to help the reader to better assess the relative proportion of both types of particles. At least, I would recommend that they change for more contrasted colors, in order to help the reader to better visually decipher between Cr(VI) and total Cr particles.  
+
+## Figure S6  
+
+I would recommend to change the letters for A, B and C in the legend to fit with the letters reported on the figure.  
+
+## Table S6  
+
+Total concentrations in bulk surface soil- ash samples (and some \(< 53\mu m\) fractions) are displayed, but total concentration in bulk soil samples across the whole soil cores are not provided. I would recommend that the authors provide these data (at least those for Cr) as mean total concentration for rhyolite, melange and serpentine soils in SI, either in the form of a Table or as a figure similar to Figure 2. Such a figure would better show that the fractions of reactive Cr(VI) is very low compared to total Cr concentration. In the same way, I would have been interested to see a figure similar to Figure 2 that would have depicted the fraction of reactive Cr(VI) as a function of the total Cr concentration. Indeed, such a figure would have helped to check if the fraction of reactive Cr(VI) is really higher in serpentine soils. Even if I agree with the authors that the concentration of reactive Cr(VI) is the most relevant parameter to assess a potential environmental and/or health risk, the fraction of reactive Cr(VI) could further inform on the actual mechanism(s) and/or soil characteristic(s) that favor Cr(III) to
+
+<--- Page Split --->
+
+Cr(VI) oxidation in burned soils. But, maybe this question is beyond the scope of the paper...  
+
+## Figure S7  
+
+Figure S7Legend : ... from a serpentine chaparral soil...Why did the authors not tried to analyze more some Cr(VI) areas on panel A ?The XANES spectrum at point 1 on panel C shows a well-marked Cr(VI) pre-edge peak but the color code indicate rather low amounts of Cr(VI) at this point. Could the authors explain that ?  
+
+## Suggested additional references  
+
+Suggested additional referencesRascio et al., 2022. Evidence of hexavalent chromium formation and changes of Cr speciation after laboratory- simulated fires of composted tannery sludges long- term amended agricultural soils. Journal of Hazardous Materials, 436, 129117. https://doi.org/10.1016/j.jhazmat.2022.129117Terzano et al., 2021. Fire effects on the distribution and bioavailability of potentially toxic elements (PTEs) in agricultural soils. Chemosphere, 130752. https://doi.org/10.1016/j.chemosphere.2021.130752Ré et al., 2021. Cytotoxic effects of wildfires ashes : In- vitro responses of skin cells. Environmental Pollution, 285, 117279. https://doi.org/10.1016/j.envpol.2021.117279Jahn et al., 2021. Metallic and crustal elements in biomass- burning aerosols and ash: Prevalence, significance, and similarities to soil particles. ACS Earth and Space Chemistry, 5, 136- 148. https://dx.doi.org/10.1021/acsearthspacechem.0c00191Xu et al., 2020. Wildfires, global climate change, and human health. The New England Journal of Medicine, 383, 2173- 2181. https://doi.org/10.1056/NEJMsr2028985
+
+<--- Page Split --->
+
+## Reviewer #2 (Remarks to the Author):  
+
+The authors have to highlight the novelty of their manuscript. The abstract should be revised to attract the reader's attention. The Introduction section should be improved by adding references dealing soil contamination issues. However, the problem is that the English and the whole organisation of the present version are definitely below an acceptable standard for an international scientific journal. Analytical quality control is missing. Detection limits of the applied methods should be reported. The main problem for this manuscript is its structure. Major parts are missing from the manuscript. My suggestion is to reject this manuscript and encourage the authors to submit a more mature manuscript.
+
+<--- Page Split --->
+
+## Response to Review Comments  
+
+## Reviewer #1  
+
+This manuscript presents the results of a study aimed at emphasizing the health threat potentially arising from chromium transformation to its most harmful \(\mathrm{Cr(VI)}\) form in soils and ashes upon wildfires. To reach their objective, the authors analyzed 38 cores drilled in more or less severely burned soils at four Preserves that experienced large wildfires in the North Coast Range of California (USA). The results obtained confirm the catalytic effect of high temperatures generated by wildfires on chromium oxidation in soil, and they point to control of soil geology and fire severity on this effect. They also interestingly indicate that reactive \(\mathrm{Cr(VI)}\) can reach dangerous levels in wind- dispersible particulates found in the surficial layers of the soil at ultramafic settings, and that this harmful form of chromium can persist in the soil/ash system up to one year after the wildfire if rainfalls are not significant. These results yield to the conclusion that more work is required to further evaluate this potential risk around the world.  
+
+I found this paper very pleasant to read. The subject is well introduced, the sites and samples are well described and fit with the objective of the study, the results are well presented and they support the discussion that gives proper attention to the existing literature on the topic and yield to the novel conclusions that geology and fire intensity are important drivers of harmful hexavalent chromium in soils and wind- dispersible soil/ash surface particles, and that the associated health risk can persist up to one year after fire.  
+
+Although the scale of the study (North Coast Range of California, USA) could at first be considered too short to extend its findings at the global scale, I agree with the authors that the number and diversity of sites studied can be considered enough to address this issue. I also agree with the concluding remark on the importance of further evaluating the potential risk of wildfire- induced harmful hexavalent chromium in wind- dispersible soil/ash particles at the global scale. Finally, I agree with the authors that the results provided in this study deserve to be shared with a large audience, ranging from scientists working on the topic to policy makers, public administrators and nature managers.  
+
+For all these reasons, I consider that this manuscript deserves to be published in Nature Communications. I have listed below few issues that I would be interested to see addressed, although only few of them are mandatory for publication.  
+
+Response: We thank the reviewer for their constructive feedback and review of the manuscript's findings and implications. We have reviewed and addressed in detail below the reviewer's suggestions.  
+
+Changes: Please see our specific changes to the reviewer's suggestions below.  
+
+Figure 1: I agree that a large fraction of ultramafic/mafic areas are concerned by wildfires at the global scale. Figure 1 shows that these areas are mainly located in the tropical region where most soils are deeply weathered (Ultisols and Oxisols, according to the USDA classification). However, the Fe contents reported in Table S6 suggest that the soils studied in this work do not
+
+<--- Page Split --->
+
+correspond to these soil types. This raises the question of the actual representativeness of the results regarding tropical areas. I would recommend that the authors comment on that point, and maybe further consider it in their concluding remarks.  
+
+Response: We agree with the reviewer that the soils in this study are not as deeply weathered as Oxisols and Ultisols, like lateritic soils, common across tropical climate regions (e.g., New Caledonia, Cuba, Brazil, Malaysia, Indonesia, Madagascar, northern Australia). In our study, the average Fe concentration across burned and unburned bulk ultramafic (serpentine) soil was 8.71 wt. \(\%\) , while lateritic soils can contain more than 5x more Fe, typically in the form of crystalline Fe oxides (e.g., hematite, goethite). Chromium is predominantly found within these crystalline Fe oxides. Additionally, lateritic soils are often depleted in Ca, Mg, and Si. The Fe concentrations of our study soils are representative of serpentine soils in Mediterranean or temperate climates (e.g., California, Oregon, Washington, Turkey, Balkans) with moderate Fe oxide content and neutral to alkaline pH. We have added a paragraph within the Discussion section that addresses the characterization of ultramafic soils globally and the limitations of soil types within our study site when considering mechanisms in lateritic and highly weathered metal- rich soils in the tropical regions. Furthermore, we place our findings from field analysis in the context of laboratory studies that examine Cr(VI) generation from Cr(III) solids common to lateritic soils.  
+
+Changes: Within the Discussion we now state:  
+
+"While recognized for urban fires, threats from metal exposure in smoke and dust need to be recognized within wildland fires arising on metal- rich geologies. Across tropical climate regions, deeply weathered lateritic soils are common, in which Cr is predominantly found within crystalline Fe oxides (e.g., hematite, goethite), and soil Fe content may exceed 50 wt. \(\%\) . Past work has quantified the Cr oxidation capacity of Cr- bearing Fe oxides and lateritic soils during heating simulations \(^{35,62}\) . For example, the greatest Cr(VI) formation (more than 40% of total Cr) upon heating hematite occurred at temperatures less than \(400^{\circ}\mathrm{C}\) , while up to 100% of total Cr in goethite transformed to Cr(VI) at \(800^{\circ}\mathrm{C}\) . In Mediterranean or temperate climates (similar to this study's region), ultramafic soils are relatively more enriched in Cr- bearing phyllosilicate minerals (e.g., serpentine), Fe oxide content is moderate (typically less than 10 wt. \(\%\) ) with more amorphous secondary phases, and soil pH is neutral to alkaline pH. In our study, and under natural wildfire conditions, we observed that up to about 0.015% of total Cr was reactive Cr(VI) in burned serpentine soils. Chromium(VI) generation during wildfires depend on fire conditions and host mineralogy; thus, the extent of Cr(VI) formation in lateritic soils may differ from temperate serpentine soils, but field observations of Cr(VI) in burned lateritic soils following wildfires are currently lacking."  
+
+Figure 3: Maybe, the authors could add the results of comparative statistics tests on barplots and boxplots (either directly report the \(p\) - values or show corresponding labels as \(***\) or \(**\) )?  
+
+Response: We performed a one- way ANOVA for Figure 3c and found no statistically significant difference between the three groups (p- value = 0.164). In Figure 3b, we are illustrate reactive concentrations within surface soil- ash based on fire severity (low
+
+<--- Page Split --->
+
+versus moderate- high) and particle size (bulk soil less than \(2\mathrm{mm}\) versus the silt and clay- sized fraction less than \(53\mu \mathrm{m}\) ). We are unable to run a two- way ANOVA because burn severity sample sizes are not equal.  
+
+Changes: For Figure 3, we now state "Reactive \(\mathrm{Cr(VI)}\) concentrations ranged from 64 to \(1,060\mu \mathrm{g / kg}\) , with a median concentration of \(257\mu \mathrm{g / kg}\) , and remained elevated compared to concentrations (5- 64 \(\mu \mathrm{g / kg}\) ) within the near surface depths (0- 2 cm) of unburned serpentine soil (Figure 3c); however, these differences were not statistically significant (p- value \(= 0.164\) )."  
+
+Do they have any hypothesis to explain the much higher concentration in reactive \(\mathrm{C(VI)}\) measured in the A7 surface soil- ash sample from the serpentine chaparral landscape? The data provided in Table S6 indicate about twice more total \(\mathrm{Cr}\) in this sample compared to sample A6 (for instance), but the concentration in reactive \(\mathrm{Cr(VI)}\) is more than 3 times higher in the bulk fraction and more than 25 three times higher in the \(< 53\mu \mathrm{m}\) fraction. Are there any mineralogical differences with the other moderately- highly burned serpentine soil- ash samples (A3- A6) that could help to explain that?  
+
+Response: At site A7, the soil experienced longer burning duration and fire intensities with greater biomass combustion that may further contribute to the high- levels of \(\mathrm{Cr(VI)}\) . Unlike A7 (original Figure S9; revised Figure S11), we did not observe mineralogical changes in bulk composition for samples A3- A6 in the surface soil and ash compared to underlying burned soil. We suspect that high temperatures did not persist for sufficient time to alter bulk mineralogy in the latter samples.  
+
+Changes: Within the Results we state: "At site A7, the average reactive \(\mathrm{Cr(VI)}\) concentration was more than three times greater than other moderate- high fire severity sites (Figure 3b). We suspect that longer burning duration and fire intensities with greater biomass combustion contributed to the relatively high- levels of \(\mathrm{Cr(VI)}\) , as this was a severely burned forested area. Importantly, ash from severely burned areas concentrate alkali (Na, K) and alkaline earth (Ca, Mg) metals (often from biomass combustion) that are key for the thermal oxidation of \(\mathrm{Cr(III)}\) \(^{13,43}\) . For example, \(\mathrm{CaCrO_4}\) was noted after agricultural soil amended with composted \(\mathrm{Cr(III)}\) - rich tannery sludge was heated at \(500^{\circ}\mathrm{C}\) \(^{55}\) ."  
+
+Figure 4: I wonder if this figure should be maintained in the main text. First, I am not convince that it really supports the assumption that "total \(\mathrm{Cr(VI)}\) was most abundant in wind- dispersible soils and ash particulates after high fire severity conditions compared to a low severity sample". Second, I do not understand why the two figures are plotted on different sizes (A is larger than B).  
+
+It could maybe be replaced by a figure similar to Figure S7 (maybe use only panels A and B or panels C and D from Figure S7 and add two similar panels from a low fire severity burn sample from a serpentine chaparral soil)?
+
+<--- Page Split --->
+
+Whatever, if Figure 4 is maintained in the main text, I would recommend that the authors add some histograms with the estimated number of Cr(VI) and total Cr particles on the two figures, in order to help the reader to better assess the relative proportion of both types of particles. At least, I would recommend that they change for more contrasted colors, in order to help the reader to better visually decipher between Cr(VI) and total Cr particles.  
+
+Response: We appreciate the reviewer's feedback regarding Figure 4. We agree that our results of greater Cr(VI)- bearing particles in high severity soil- ash versus low- severity conditions can be clearer, especially by including a histogram of particles containing Cr(VI). Regarding the reviewer's concern about the different size plots, the sample area for large- scale micro- XRF maps of each thin section were not held constant during data collection, resulting in different mapped areas. Due to limitations of the analysis software, we are unable to change for more contrasted colors, but we can update the min/max values for color brightness.  
+
+We have revised Figure 4 to focus on a \(1 - \mu \mathrm{m}\) resolution XRF image containing particles from a high fire severity sample with and without measurable Cr(VI) by XANES analysis. We then moved the original Figure 4 to the Supplementary Information as Figure S8, and added another example of Cr(VI)- containing particles in a \(1 - \mu \mathrm{m}\) resolution XRF image as Figure S9. We also revised the main text related to the figures to highlight the presence of Cr(VI)- containing particles within the high fire severity samples, which was not apparent in low fire severity samples.  
+
+Changes: See revised Figure 4, Figure S8, and Figure S9 below.  
+
+"Using micro- scale X- ray techniques, Cr(VI)- containing soil and ash particulates were identified in a high fire severity sample as opposed to a low fire severity sample (Figure 4; Figure S8). Here, Cr(VI) was associated with mineral surfaces (e.g., adsorbed) or enriched in relatively low- Cr particles with Ca and K (Figure 4, Figure S9). Consistent with particle analysis, reactive Cr(VI) concentrations spanned from 326 to \(13,000 \mu \mathrm{g / kg}\) (Figure 3b)."
+
+<--- Page Split --->
+![](images/Figure_unknown_0.jpg)
+
+<center>Figure 4 | Total Cr(VI) in wind-dispersible soil and ash particles. a. \(\mu\) -XRF image (pixel resolution: \(1 \mu m\) ) showing the relative intensity of Cr(VI) (green; estimated as the intensity ratio at 5993 and 6010 eV) and total Cr (blue; measured at 6010 eV) within the \(< 53 - \mu m\) size fraction of Cr-bearing soil-ash particulates from a serpentine chaparral that experienced high fire severity (A7). b. Normalized \(\mu\) -XANES spectra (Cr K-edge) from numbered locations on Cr-bearing particles in a. Dashed lines indicate energies characteristic of Cr(VI) (5993 eV), Cr(III) (6003 eV), and total Cr (6010 eV), at which \(\mu\) -XRF images were also collected. </center>
+
+<--- Page Split --->
+
+Figure S8 | Total Cr(VI) in wind-dispersible soil and ash particles. \(\mu\) - XRF image showing particle distribution of total Cr(VI) (green; estimated as the intensity ratio at 5993 and 6010 eV) and total Cr (blue; measured at 6010 eV) within the \(< 53 - \mu \mathrm{m}\) size fraction of Cr-bearing soil- ash particulates from a. high fire severity site (A7) and b. low fire severity site (A1) in a serpentine chaparral.  
+
+![](images/Figure_unknown_1.jpg)
+
+<center>Figure S9 | Total Cr(VI) in wind-dispersible soil and ash particles. a. \(\mu\) -XRF image (pixel resolution: \(1\mu \mathrm{m}\) ) showing the relative intensity of \(\mathrm{Cr(VI)}\) (green; estimated as the intensity ratio at 5993 and \(6010\mathrm{eV}\) ) and total Cr (blue; measured at \(6010\mathrm{eV}\) ) within the \(< 53 - \mu \mathrm{m}\) size fraction of Cr-bearing soil-ash particulates from a serpentine chaparral that experienced high fire severity (A7). b. Normalized \(\mu\) -XANES spectra (Cr K-edge) from numbered locations on Cr-bearing particles in a. Dashed lines indicate energies characteristic of \(\mathrm{Cr(VI)}\) (5993 eV), \(\mathrm{Cr(III)}\) (6003 eV), and total Cr (6010 eV). </center>  
+
+![](images/Figure_4.jpg)
+
+
+<--- Page Split --->
+
+Figure S6: I would recommend to change the letters for A, B and C in the legend to fit with the letters reported on the figure.  
+
+Response: Based on the reviewer's recommendation, we have revised the figure, accordingly, by changing uppercase letters to lowercase, in addition to the other figures with sub- panels (similarly identified by Reviewer #2).  
+
+Changes: We changed the letters to lowercase, as requested.  
+
+Table S6: Total concentrations in bulk surface soil- ash samples (and some \(< 53 \mu \mathrm{m}\) fractions) are displayed, but total concentration in bulk soil samples across the whole soil cores are not provided. I would recommend that the authors provide these data (at least those for Cr) as mean total concentration for rhyolite, mélange and serpentine soils in SI, either in the form of a Table or as a figure similar to Figure 2. Such a figure would better show that the fractions of reactive \(\mathrm{Cr(VI)}\) is very low compared to total Cr concentration.  
+
+Response: We thank the reviewer for their recommendation, which we have addressed in the revision. Please refer to our response and changes to the reviewer's next comment related to including a figure showing the fraction of reactive \(\mathrm{Cr(VI)}\) to total Cr concentrations.  
+
+Changes: Using total Cr concentrations reported for soil cores in Table S1, we have revised Table S6 to include the mean total element concentrations for Cr, Fe, Mn, Ni, Ca, Mg, Na, and K in addition to the surface soil- ash sample data so that the reader can compare elemental concentrations in surface soil and ash to bulk soil from different geologies.
+
+<--- Page Split --->
+
+Table S6 | Physicochemical characteristics of bulk soil and ash (up to 2 mm), and selected fine size fractions less than 53 μm, collected from surface layers of the burned serpentine chaparral, and mean elemental concentrations from bulk underlying soil based on geology type (rhyolitic, mélange, and serpentine).
+
+<table><tr><td>ID</td><td>Fire<br>Severitya</td><td>% Sandb<br>(2-0.05 mm)</td><td>% Siltc<br>(53-2 μm)</td><td>% Clayc<br)&lt;2 μm)</td><td>Cr<br>mg/kg</td><td>Fe<br>mg/g</td><td>Mn<br>mg/kg</td><td>Ni<br>mg/kg</td><td>Ca<br>mg/g</td><td>Mg<br>mg/g</td><td>Na<br>mg/g</td><td>K<br>mg/g</td></tr><tr><td>Surface Soil-Ash</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td>A1</td><td>L</td><td></td><td></td><td></td><td>1147</td><td>64.7</td><td>1351</td><td>1528</td><td>15.5</td><td>77.3</td><td>4.63</td><td>3.76</td></tr><tr><td>A2</td><td>L</td><td></td><td></td><td></td><td>1532</td><td>69.1</td><td>1203</td><td>2530</td><td>10.2</td><td>158</td><td>&lt;0.1</td><td>1.65</td></tr><tr><td>A3</td><td>M/H</td><td></td><td></td><td></td><td>1606</td><td>84.2</td><td>1480</td><td>3117</td><td>9.3</td><td>148</td><td>&lt;0.1</td><td>1.54</td></tr><tr><td>A4</td><td>M/H</td><td>87.5</td><td>11.7</td><td>0.8</td><td>2256</td><td>78.9</td><td>1438</td><td>3380</td><td>8.7</td><td>159</td><td>&lt;0.1</td><td>0.90</td></tr><tr><td>A5</td><td>M/H</td><td></td><td></td><td></td><td>999</td><td>57.7</td><td>1022</td><td>1849</td><td>10.2</td><td>125</td><td>2.10</td><td>2.48</td></tr><tr><td>A6</td><td>M/H</td><td>78.2</td><td>20.2</td><td>1.6</td><td>1970</td><td>79.2</td><td>1555</td><td>2726</td><td>29.9</td><td>174</td><td>&lt;0.1</td><td>2.86</td></tr><tr><td>A7</td><td>M/H</td><td>85.7</td><td>12.9</td><td>1.4</td><td>4829</td><td>102</td><td>1543</td><td>2643</td><td>17.9</td><td>211</td><td>&lt;0.1</td><td>2.29</td></tr><tr><td>Less than 53 μm size fraction</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td>A4</td><td>M/H</td><td></td><td></td><td></td><td>1133</td><td>94.7</td><td>1975</td><td>3042</td><td>41.5</td><td>124</td><td>&lt;0.1</td><td>3.38</td></tr><tr><td>A6</td><td>M/H</td><td></td><td></td><td></td><td>946</td><td>82.7</td><td>1910</td><td>2294</td><td>44.2</td><td>133</td><td>&lt;0.1</td><td>4.41</td></tr><tr><td>A7</td><td>M/H</td><td></td><td></td><td></td><td>1643</td><td>105</td><td>2547</td><td>3489</td><td>34.0</td><td>189</td><td>2.02</td><td>3.91</td></tr><tr><td>Bulk Soild</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td>Rhyolitic \((n=7)\)</td><td></td><td></td><td></td><td></td><td>162</td><td>34.5</td><td>796</td><td>88</td><td>10.7</td><td>7.9</td><td>13.6</td><td>9.70</td></tr><tr><td>Melange \((n=16)\)</td><td></td><td></td><td></td><td></td><td>314</td><td>50.0</td><td>855</td><td>259</td><td>10.7</td><td>37.3</td><td>8.17</td><td>13.2</td></tr><tr><td>Serpentine \((n=10)\)</td><td></td><td></td><td></td><td></td><td>2373</td><td>87.1</td><td>1511</td><td>2929</td><td>4.4</td><td>150</td><td>3.61</td><td>1.69</td></tr></table>
+
+a L = Low severity, M = Moderate severity, H = High severity
+
+b Determined by sieve analysis
+
+c Determined by laser diffraction particle size counter
+
+d Bulk soil concentrations are mean values using all soil cores (fire-affected and unburned) for each geology type:rhyolitic \((n=7)\) , mélange \((n=16)\) , and serpentine \((n=10)\) .
+
+<--- Page Split --->
+
+In the same way, I would have been interested to see a figure similar to Figure 2 that would have depicted the fraction of reactive \(\mathrm{Cr(VI)}\) as a function of the total \(\mathrm{Cr}\) concentration. Indeed, such a figure would have helped to check if the fraction of reactive \(\mathrm{Cr(VI)}\) is really higher in serpentine soils. Even if I agree with the authors that the concentration of reactive \(\mathrm{Cr(VI)}\) is the most relevant parameter to assess a potential environmental and/or health risk, the fraction of reactive \(\mathrm{Cr(VI)}\) could further inform on the actual mechanism(s) and/or soil characteristic(s) that favor \(\mathrm{Cr(III)}\) to \(\mathrm{Cr(VI)}\) oxidation in burned soils. But, maybe this question is beyond the scope of the paper...  
+
+Response: We thank the reviewer for their recommendation. We agree that the reactive \(\mathrm{Cr(VI)}\) fraction of total \(\mathrm{Cr}\) in soil and soil- ash is low; however, the fraction is relatively higher in near surface soil depths compared to past studies quantifying natural \(\mathrm{Cr(III)}\) oxidation in unburned soils, including a 2017 study at McLaughlin Natural Reserve. Moreover, as the reviewer notes, the hazard imposed by the particulates is related to the reactive \(\mathrm{Cr(VI)}\) . The percentage of total \(\mathrm{Cr}\) that was reactive \(\mathrm{Cr(VI)}\) in unburned serpentine soil was consistent with previous measurements (McClain et al., 2017). Interestingly, the reactive \(\mathrm{Cr(VI)}\) fraction differs based on geology (rhyolite, mélange, and serpentinite). The reactive \(\mathrm{Cr(VI)}\) fraction in rhyolitic and mélange soils composed more of the total \(\mathrm{Cr}\) content than the relative fraction within serpentine soils. In order to highlight these variations across soil depth in fire- affected and unburned sites, we added a figure to the supplementary information.  
+
+Changes: Figure S6 (below) was added to the Supplementary Information. Succeeding figure numbers were updated based on this addition. Within the Results and Discussion sections, we now state:  
+
+"Average \(\mathrm{Cr(VI)}\) concentrations generated in soils derived from mélange (Figure 2b) were more than double the respective levels in rhyolitic soil (Figure 2a) regardless of overlapping ranges in total \(\mathrm{Cr}\) content, 152- 954 and 102- 338 mg/kg, respectively, reflecting the potential contribution of differing mineralogy to \(\mathrm{Cr(VI)}\) generation (Figure S6) 35. "  
+
+"In our study, and under natural wildfire conditions, we observed that up to about \(0.015\%\) of total \(\mathrm{Cr}\) was reactive \(\mathrm{Cr(VI)}\) in burned serpentine soils (Figure S6)."
+
+<--- Page Split --->
+
+Figure S6 | Fraction of total Cr that is reactive \(\mathrm{Cr(VI)}\) in burned and unburned soils. The ratio of reactive \(\mathrm{Cr(VI)}\) to total Cr concentrations (as a percentage) within a. rhyolite- , b. mélange- , c. serpentinite- derived soil profiles (0- 16 cm) that were not burned (gray; rhyolite, \(n =\) 3; mélange, \(n = 7\) ; serpentinite, \(n = 3\) ) or were fire- affected (colored; rhyolite, \(n = 4\) ; mélange, \(n\) \(= 9\) ; serpentinite, \(n = 7\) ). Percentages were also plotted for serpentinite- derived soil (0- 20 cm) from McClain et al. (2017) in c. for comparison. Each point represents the average percentage for a soil core based on triplicate measurements.  
+
+![PLACEHOLDER_14_0]
+
+<center>Figure S7: Legend : ... from a serpentine chaparral soil... </center>  
+
+Response: We agree with the reviewer's recommendation and have made the appropriate changes. Based on the reviewer's previous recommendations, this figure has revised as Figure 4 and Figure S9.  
+
+Changes: (Figure 4 and Figure S9 caption) "... particles ( \(< 53 \mu \mathrm{m}\) ) from a serpentine chaparral that experienced high fire severity ..."  
+
+Why did the authors not tried to analyze more some \(\mathrm{Cr(VI)}\) areas on panel A? The XANES spectrum at point 1 on panel C shows a well- marked \(\mathrm{Cr(VI)}\) pre- edge peak but the color code indicate rather low amounts of \(\mathrm{Cr(VI)}\) at this point. Could the authors explain that?  
+
+Response: Our XANES analysis was used to corroborate our bulk measurements of reactive \(\mathrm{Cr(VI)}\) . Particles within the XRF map were used to denote the presence and abundance of \(\mathrm{Cr(VI)}\) , corroborating (and visualizing) the reactive fraction measurements.  
+
+Changes: "To corroborate bulk measurements of reactive \(\mathrm{Cr(VI)}\) , we combined multi- energy mapping with \(\mu\) - XANES of select spots on particles to confirm the presence of \(\mathrm{Cr(VI)}\) ."  
+
+## Suggested additional references  
+
+Rascio et al., 2022. Evidence of hexavalent chromium formation and changes of \(\mathrm{Cr}\) speciation after laboratory- simulated fires of composted tannery sludges long- term amended agricultural
+
+<--- Page Split --->
+
+soils. Journal of Hazardous Materials, 436, 129117. https://doi.org/10.1016/j.jhazmat.2022.129117 Terzano et al., 2021. Fire effects on the distribution and bioavailability of potentially toxic elements (PTEs) in agricultural soils. Chemosphere, 130752. https://doi.org/10.1016/j.chemosphere.2021.130752 Ré et al., 2021. Cytotoxic effects of wildfires ashes : In- vitro responses of skin cells. Environmental Pollution, 285, 117279. https://doi.org/10.1016/j.envpol.2021.117279 Jahn et al., 2021. Metallic and crustal elements in biomass- burning aerosols and ash: Prevalence, significance, and similarities to soil particles. ACS Earth and Space Chemistry, 5, 136- 148. https://dx.doi.org/10.1021/acsearthspacechem.0c00191 Xu et al., 2020. Wildfires, global climate change, and human health. The New England Journal of Medicine, 383, 2173- 2181. https://doi.org/10.1056/NEJMsr2028985  
+
+Response: We thank the reviewer for sharing additional references. We agree that these references are relevant to the study.  
+
+Changes: We have added these references suggested above, in addition to a few other recently published and relevant studies, to discussions within the main text.  
+
+"Global wildfire activity represents a rising distributed health risk from smoke and dust inhalation \(^{5 - 10}\) ."  
+
+10. Xu, R. et al. Wildfires, Global Climate Change, and Human Health. N. Engl. J. Med. 383, 2173-2181 (2020).  
+
+"Increased heavy metals in PM have been documented during wildfire episodes and may induce cytotoxicity, increase lung cancer risks, and greatly contribute to oxidative stress \(^{19 - 30}\) ."  
+
+"Suburban fires illustrates the impacts of inhaling Cr(VI)- containing ash within the respiratory tract by measuring Cr(VI) leached with a simulated lung fluid \(^{36,37}\) and discerning Cr mineralogy within nano- sized particulates (< 100 nm) \(^{22,30}\) ."  
+
+21. Boaggio, K. et al. Beyond Particulate Matter Mass: Heightened Levels of Lead and Other Pollutants Associated with Destructive Fire Events in California. Environ. Sci. Technol. 56, 14272-14283 (2022).  
+
+22. Alshehri, T. et al. Wildland-urban interface fire ashes as a major source of incidental nanomaterials. J. Hazard. Mater. 443, 130311 (2023).  
+
+28. Jahn, L. G. et al. Metallic and crustal elements in biomass-burning aerosol and ash: Prevalence, significance, and similarity to soil particles. ACS Earth Sp. Chem. 5, 136-148 (2021).  
+
+29. Ré, A. et al. Cytotoxic effects of wildfire ashes: In-vitro responses of skin cells. Environ. Pollut. 285, 117279 (2021).
+
+<--- Page Split --->
+
+"Metals in soils and ash are commonly linked to structural burning within wildland- urban interfaces (WUI) \(^{1,31 - 33}\) , with negligible awareness of wildland landscapes (soils and ash) as an alternative and highly distributed source \(^{1}\) ."  
+
+32. Alexakis, D. E. Suburban areas in flames: Dispersion of potentially toxic elements from burned vegetation and buildings. Estimation of the associated ecological and human health risk. Environ. Res. 183, 109153 (2020).  
+
+33. Alam, M. et al. Identification and quantification of Cr, Cu, and As incidental nanomaterials derived from CCA-treated wood in wildland-urban interface fire ashes. J. Hazard. Mater. 445, 130608 (2023).  
+
+"Following wildfires, severely burned areas are often barren and blanketed with ash and loose, rough topsoil leading to enhanced post-fire wind and water erosion \(^{8,39 - 42}\) ."  
+
+41. Yu, Y. & Ginoux, P. Enhanced dust emission following large wildfires due to vegetation disturbance. Nat. Geosci. 2022 1511 15, 878-884 (2022).  
+
+42. Shakesby, R. A. & Doerr, S. H. Wildfire as a hydrological and geomorphological agent. Earth-Science Rev. 74, 269-307 (2006).  
+
+"For example, CaCrO4 was noted after agricultural soil amended with composted Cr(III)- rich tannery sludge was heated at \(500^{\circ}\mathrm{C}^{55}\) ."  
+
+55. Rascio, I. et al. Evidence of hexavalent chromium formation and changes of Cr speciation after laboratory-simulated fires of composted tannery sludges long-term amended agricultural soils. J. Hazard. Mater. 436, 129117 (2022).
+
+<--- Page Split --->
+
+## Reviewer 2  
+
+The authors have to highlight the novelty of their manuscript. The abstract should be revised to attract the reader's attention. The Introduction section should be improved by adding references dealing soil contamination issues. However, the problem is that the English and the whole organisation of the present version are definitely below an acceptable standard for an international scientific journal. Analytical quality control is missing. Detection limits of the applied methods should be reported. The main problem for this manuscript is its structure. Major parts are missing from the manuscript. My suggestion is to reject this manuscript and encourage the authors to submit a more mature manuscript.  
+
+Response: We appreciate the reviewer's feedback and have sought to make the abstract, and the manuscript, have more pizzazz. With that said, the manuscript was formatted specifically to the Nature guidelines. Further, the authors are all native speakers, and the senior author has published several hundred articles, including ones in Nature and Science. The present manuscript holds to those same standards. It is also worth noting that counter to Reviewer 2, Reviewer 1 stated "I found this paper very pleasant to read. The subject is well introduced, the sites and samples are well described and fit with the objective of the study, the results are well presented and they support the discussion...". Further, several established authors at Stanford have read the manuscript and all support the writing and presentation. Thus, while we don't want to dismiss the comments of the Reviewer, we do see them as anonymous in comparison to others.  
+
+Based on the reviewer's feedback, we have revised our Methods section to include more information regarding our analyses. We have added detection limits for \(\mathrm{Cr(VI)}\) and \(\mathrm{NH_4}\) measurements, and have revised related Figures and Tables to reflect these changes. With the exception of Na (detection limit of \(0.1\mathrm{mg / g}\) ), total concentrations for all elements reported using XRF were significantly greater than respective detection limits. We periodically analyzed certified reference material, NIST 2711a, with bulk soil samples to confirm accuracy of the XRF instrument. For aqueous extractions and associated instrument analyses, we tracked quality assurance in multiple ways. For each round of aqueous extractions using \(10\mathrm{mM}\mathrm{K_2HPO_4 / KH_2PO_4}\) solution, we included at least two centrifuge tubes containing the phosphate buffer solution and no soil/ash that were analyzed similar to samples for \(\mathrm{Cr(VI)}\) and \(\mathrm{NH_4}\) . Unless soil mass was limited, extractions were conducted in triplicate to assess sample heterogeneity. On the UV- Vis and ICP- MS, we analyzed instrument blanks every 15- 20 samples and multiple quality control standard solutions prepared with certified Cr reference solutions throughout each analysis.  
+
+Changes: We have modified the Abstract, added references to the Introduction section, and have sought to ensure the format and writing are consistent with the expected quality of the Nature journals. We have also revised the Methods to include detection limits, where relevant, and have similarly revised Figures and Tables to reflect non- detectable sample concentrations throughout the manuscript.  
+
+Within the Methods section, we now state:
+
+<--- Page Split --->
+
+"Aqueous Extractions and Chemical AnalysisReactive \(\mathrm{Cr(VI)}\) concentrations (most available fraction, including dissolved and adsorbed \(\mathrm{Cr(VI)}\) ) in bulk soil ash samples (bulk and particle size fraction \(< 53 \mu \mathrm{m}\) ) and within soil cores were extracted with \(10 \mathrm{mM} \mathrm{K}_2\mathrm{HPO}_4 / \mathrm{KH}_2\mathrm{PO}_4\) (buffered at \(\mathrm{pH} 7.2\) ) \(^{79,80}\) . Phosphate effectively competes with \(\mathrm{Cr(VI)}\) ions for surface adsorption sites. At circumnatural to alkaline pH ranges in natural soils, it's expected that nearly all aqueous \(\mathrm{Cr}\) is present in the hexavalent form, and that \(\mathrm{Cr(VI)}\) concentrations will be primarily limited by adsorption \(^{81}\) . The clay size fraction (less than \(2 - \mu \mathrm{m}\) diameter) typically has a dominant influence on species retention given their high surface areas and greater number of adsorption sites; therefore, it is likely that the reactive \(\mathrm{Cr(VI)}\) concentrations measured here largely represent the fraction of \(\mathrm{Cr(VI)}\) associated with clay particles. Triplicate samples were agitated in a 1:4 soil/solution ratio for \(24 \mathrm{h}\) , centrifuged (30 min, \(4000 \mathrm{rpm}\) , \(4^{\circ}\mathrm{C}\) ), and filtered through \(0.22 - \mu \mathrm{m}\) filters. A subsample of unacidified filtrate was used to quantify aqueous \(\mathrm{Cr(VI)}\) concentrations using the diphenylcarbazide (DPC) method on a UV- Vis spectrophotometer (Shimadzu UV- 1601) \(^{79,82}\) . The detection limit was \(3 \mu \mathrm{g / L}\) (approximately \(12 \mu \mathrm{g / kg}\) ) \(^{82}\) . Total \(\mathrm{Cr}\) concentrations were determined with inductively coupled plasma mass spectrometry (ICP- MS, Thermo Scientific XSERIES 2), and confirmed that approximately all aqueous \(\mathrm{Cr}\) was in the form of \(\mathrm{Cr(VI)}\) in unburned soil and burned soil and ash, similarly observed in previous studies \(^{36,37}\) . An aliquot of each soil extract was immediately acidified post- filtration and stored in \(2\%\) nitric acid at \(4^{\circ}\mathrm{C}\) until ICP- MS analysis.  
+
+To determine relative differences in \(\mathrm{K}^+\) - extractable \(\mathrm{NH_4^+}\) concentrations (mg \(\mathrm{NH_4^+ - N / kg)}\) within burned and unburned soils (Figure S10), additional unacidified samples (after \(\mathrm{K_2HPO_4 / KH_2PO_4}\) extraction) from 30 of the 38 total soil cores (21 fire- affected and 9 unburned soil cores) were frozen at \(- 20^{\circ}\mathrm{C}\) until chemical analysis. Ammonium is a direct combustion product and will be elevated in the near surface soil after wildfires depending on burn severity \(^{40}\) . Ammonium concentrations in the top 6- cm were measured in triplicate (when sample volume allowed) using a flow injection analyzer (Westco SmartChem 200 Discrete Analyzer), with a detection limit of \(0.05 \mathrm{mg / L}\) \(^{82}\) .  
+
+## Statistical Analyses  
+
+Means and standard errors were calculated for aqueous and solid- phase chemical measurements in all cores using replicates described below. Half the detection limit was used when measured concentrations were below detection limits. Total elemental concentrations were measured in 3- 4 solid- phase aliquots from each soil core (Table S1). At each soil depth interval (1- cm from 0- 6 cm; 2- cm from 6- 16 cm), triplicate aqueous extractions were conducted to evaluate reactive \(\mathrm{Cr(VI)}\) and exchangeable \(\mathrm{NH_4^+}\) concentrations (Figure 2, Figure 3a, Figure S10). In select soil depths within cores, replicates were limited (less than 3) due to solid mass or post- extraction aqueous volume.  
+
+To assess data normality, we applied the Shapiro- Wilk test and reported \(W\) statistics and p- values (Table S7). If data met normality assumptions at the \(95\%\) confidence interval (p- value \(= 0.05\) ), we used two- sided parametric tests; otherwise, we utilized two- sided nonparametric tests. Likewise, we used the f- test to determine equal variance. Unpaired \(t\)
+
+<--- Page Split --->
+
+tests were used to compare mean reactive Cr(VI) concentrations at the \(95\%\) confidence interval in near surface soil (0- 2 cm) of fire- affected and unburned sites based on geology. If one or both datasets were not normally distributed, such as in burned and unburned soils at control depths (10- 16 cm), Mann- Whitney U test was used. Within a soil core, we compared mean reactive Cr(VI) concentrations in surface soil (0- 2 cm) versus control depths (10- 16 cm) using either paired \(t\) test or Wilcoxon signed rank test. Detailed information about and results for each statistical analysis is provided in Tables S2- S5 and Table S7 the Supplementary Information. All statistical analyses were executed using the stats package in R (v. 4.1.3)."  
+
+## From comments made on the manuscript pdf  
+
+Introduction section needs a short paragraph at the beginning to discuss elements distribution issues in wildfire impacted areas worldwide. More papers related to this paragraph will be beneficial for the paper.  
+
+May I suggest, among others, the following articles, e.g.:  
+
+1) Wildfire effects on soil quality. Application on a suburban area of West Attica (Greece).  
+
+Geosciences Journal, 25 (2), 243- 253 (https://doi.org/10.1007/s12303- 020- 0011- 1).  
+
+2) Suburban areas in flames: Dispersion of potentially toxic elements from burned vegetation and buildings. Estimation of the associated ecological and human health risk. Environmental Research, 183, 109153, https://doi.org/10.1016/j.envres.2020.109153.  
+
+Research, 183, 109153, https://doi.org/10.1016/j.envres.2020.109153.  
+
+3) Elements' Content in Stream Sediment and Wildfire Ash of Suburban Areas in West Attica (Greece). Water 2022, 14, 310. https://doi.org/10.3390/w14030310  
+
+Response: We thank the reviewer for the suggestion to add a paragraph discussing elemental concentrations of soil and ash after wildfires. We address past work that has quantified elemental concentrations in soils and ash in the first two paragraphs of the Introduction. We have added the second reference the reviewer suggested to our main text, in addition to a few recent studies on metals and their prevalence as a function of structural burning. References 1 and 3 address elemental concentrations within stream sediments of WUIs, and are ancillary to the introduction on metals within airborne particulate matter, and soil and ash, which are the focus of our study.  
+
+Changes: Based on both reviewers' recommendations, we have added the following references within the main text:  
+
+10. Xu, R. et al. Wildfires, Global Climate Change, and Human Health. N. Engl. J. Med. 383, 2173-2181 (2020).  
+
+26. Jahn, L. G. et al. Metallic and crustal elements in biomass-burning aerosol and ash: Prevalence, significance, and similarity to soil particles. ACS Earth Sp. Chem. 5, 136-148 (2021).
+
+<--- Page Split --->
+
+27. Ré, A. et al. Cytotoxic effects of wildfire ashes: In-vitro responses of skin cells. Environ. Pollut. 285, 117279 (2021).  
+
+29. Boaggio, K. et al. Beyond Particulate Matter Mass: Heightened Levels of Lead and Other Pollutants Associated with Destructive Fire Events in California. Environ. Sci. Technol. 56, 14272-14283 (2022).  
+
+30. Alshehri, T. et al. Wildland-urban interface fire ashes as a major source of incidental nanomaterials. J. Hazard. Mater. 443, 130311 (2023).  
+
+32. Alexakis, D. E. Suburban areas in flames: Dispersion of potentially toxic elements from burned vegetation and buildings. Estimation of the associated ecological and human health risk. Environ. Res. 183, 109153 (2020).  
+
+33. Alam, M. et al. Identification and quantification of Cr, Cu, and As incidental nanomaterials derived from CCA-treated wood in wildland-urban interface fire ashes. J. Hazard. Mater. 445, 130608 (2023).  
+
+41. Yu, Y. & Ginoux, P. Enhanced dust emission following large wildfires due to vegetation disturbance. Nat. Geosci. 2022 1511 15, 878-884 (2022).  
+
+42. Shakesby, R. A. & Doerr, S. H. Wildfire as a hydrological and geomorphological agent. Earth-Science Rev. 74, 269-307 (2006).  
+
+55. Rascio, I. et al. Evidence of hexavalent chromium formation and changes of Cr speciation after laboratory-simulated fires of composted tannery sludges long-term amended agricultural soils. J. Hazard. Mater. 436, 129117 (2022).  
+
+80. James, B. R. & Bartlett, R. J. Behavior of Chromium in Soils: VII. Adsorption and Reduction of Hexavalent Forms1. J. Environ. Qual. 12, 177 (1983).  
+
+81. Rai, D., Eary, L. E. & Zachara, J. M. Environmental chemistry of chromium. Sci. Total Environ. 86, 15-23 (1989).  
+
+82. McClain, C. N., Fendorf, S., Johnson, S. T., Menendez, A. & Maher, K. Lithologic and redox controls on hexavalent chromium in vadose zone sediments of California's Central Valley. Geochim. Cosmochim. Acta 265, 478-494 (2019).  
+
+Figure 2: Change A, B, and C in Figure plot panels to a, b, and c, reflected in the figure caption  
+
+Response: We appreciate the reviewer's feedback and have made the salient changes.  
+
+Changes: We have revised the figure panels (see below) in addition to references to the figure throughout the main text.
+
+<--- Page Split --->
+![PLACEHOLDER_21_0]
+  
+
+Please avoid using "we". Apply this type of revision all over the manuscript.  
+
+Response: We appreciate the reviewer's concern for the use of "we" within the manuscript. According to the editor and style guidelines for Nature Communications, the use of "we" is allowed and in some places encouraged.  
+
+Changes: No changes were made.  
+
+Please avoid using "we".  
+
+Response: Please refer to previous response.  
+
+Changes: No changes were made.  
+
+Figure 3: Change A, B, and C in Figure plot panels to a, b, and c, reflected in the figure caption  
+
+Response: We appreciate the reviewer's feedback agree with the suggestions.  
+
+Changes: We have revised the figure panels (see below) in addition to references to the figure throughout the main text.
+
+<--- Page Split --->
+![PLACEHOLDER_22_0]
+
+<center>Figure 4: Change A and B in Figure plot panels to a and b reflected in the figure caption </center>  
+
+Response: We appreciate the reviewer's feedback and agree with the suggestions.  
+
+Changes: We have revised the figure panels (see below) in addition to references to the figure throughout the main text. Please note that this figure was moved to the Supplementary Information (revised Figure S8), based on the recommendation of Reviewer 1.  
+
+![PLACEHOLDER_22_1]
+  
+
+Where are the sections "Results and Discussion" and "Conclusions"?
+
+<--- Page Split --->
+
+Response: We have followed the format for Nature, which often do not have specific sections denoted as "Results and Discussion" nor "Conclusions". Further, within the author guidelines for Nature Communication submissions: "Nature Communications is flexible with regard to the format of initial submissions. Within reason, style and length will not directly influence consideration of a manuscript. We also do not require a particular structure or format at first submission. If and when revisions are required, the editor will provide detailed formatting instructions at that time." Based on the editor's instructions for manuscript revisions, we have defined "Introduction" and "Results" sections in the manuscript according to the Nature Communications formatting instructions.  
+
+Changes: Based on the editor's instructions for manuscript revisions, we have defined "Introduction" and "Results" sections in the manuscript according to the Nature Communications formatting instructions.
+
+<--- Page Split --->

peer_reviews/12438c2ece72c0a7ec2a093dc88eaba15e637ea325a5525049bbab0bfceb822d/supplementary_0_Peer Review File/supplementary_0_Peer Review File_det.mmd

@@ -0,0 +1,523 @@
+<|ref|>title<|/ref|><|det|>[[61, 40, 506, 90]]<|/det|>
+# nature portfolio  
+
+<|ref|>title<|/ref|><|det|>[[68, 110, 361, 139]]<|/det|>
+# Peer Review File  
+
+<|ref|>title<|/ref|><|det|>[[83, 154, 914, 210]]<|/det|>
+# Metal Toxin Threat in Wildland Fires Determined by Geology and Fire Severity  
+
+<|ref|>image<|/ref|><|det|>[[57, 732, 240, 783]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[250, 732, 911, 784]]<|/det|>
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[401, 10, 599, 25]]<|/det|>
+# REVIEWER COMMENTS  
+
+<|ref|>text<|/ref|><|det|>[[373, 37, 629, 51]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>title<|/ref|><|det|>[[179, 82, 817, 100]]<|/det|>
+# Review of manuscript 395109 submitted to Nature Communication  
+
+<|ref|>title<|/ref|><|det|>[[140, 115, 856, 133]]<|/det|>
+# Metal toxin threat in wildland fires determined by geology and fire severity  
+
+<|ref|>text<|/ref|><|det|>[[241, 147, 815, 164]]<|/det|>
+Alandra Marie Lopez, Juan Lezama Pacheco and Scott Fendorf  
+
+<|ref|>text<|/ref|><|det|>[[117, 196, 881, 394]]<|/det|>
+This manuscript presents the results of a study aimed at emphasizing the health threat potentially arising from chromium transformation to its most harmful \(\mathrm{Cr(VI)}\) form in soils and ashes upon wildfires. To reach their objective, the authors analyzed 38 cores drilled in more or less severely burned soils at four Preserves that experienced large wildfires in the North Coast Range of California (USA). The results obtained confirm the catalytic effect of high temperatures generated by wildfires on chromium oxidation in soil, and they point to control of soil geology and fire severity on this effect. They also interestingly indicate that reactive \(\mathrm{Cr(VI)}\) can reach dangerous levels in wind- dispersible particulates found in the surficial layers of the soil at ultramafic settings, and that this harmful form of chromium can persist in the soil/ash system up to one year after the wildfire if rainfalls are not significant. These results yield to the conclusion that more work is required to further evaluate this potential risk around the world.  
+
+<|ref|>text<|/ref|><|det|>[[117, 410, 881, 526]]<|/det|>
+I found this paper very pleasant to read. The subject is well introduced, the sites and samples are well described and fit with the objective of the study, the results are well presented and they support the discussion that gives proper attention to the existing literature on the topic and yield to the novel conclusions that geology and fire intensity are important drivers of harmful hexavalent chromium in soils and wind- dispersible soil/ash surface particles, and that the associated health risk can persist up to one year after fire.  
+
+<|ref|>text<|/ref|><|det|>[[117, 542, 881, 689]]<|/det|>
+Although the scale of the study (North Coast Range of California, USA) could at first be considered too short to extend its findings at the global scale, I agree with the authors that the number and diversity of sites studied can be considered enough to address this issue. I also agree with the concluding remark on the importance of further evaluating the potential risk of wildfire- induced harmful hexavalent chromium in wind- dispersible soil/ash particles at the global scale. Finally, I agree with the authors that the results provided in this study deserve to be shared with a large audience, ranging from scientists working on the topic to policy makers, public administrators and nature managers.  
+
+<|ref|>text<|/ref|><|det|>[[117, 704, 880, 772]]<|/det|>
+For all these reasons, I consider that this manuscript deserves to be published in Nature Communications. I have listed below few issues that I would be interested to see addressed, although only few of them are mandatory for publication.  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 788, 192, 803]]<|/det|>
+## Figure 1  
+
+<|ref|>text<|/ref|><|det|>[[117, 807, 881, 906]]<|/det|>
+I agree that a large fraction of ultramafic/mafic areas are concerned by wildfires at the global scale. Figure 1 shows that these areas are mainly located in the tropical region where most soils are deeply weathered (Utilisols and Oxisols, according to the USDA classification). However, the Fe contents reported in Table S6 suggest that the soils studied in this work do not correspond to these soil types. This raises the question of the actual representativeness of the results regarding tropical areas. I would
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[116, 83, 880, 116]]<|/det|>
+recommend that the authors comment on that point, and maybe further consider it in their concluding remarks.  
+
+<|ref|>sub_title<|/ref|><|det|>[[117, 133, 195, 149]]<|/det|>
+## Figure 3  
+
+<|ref|>text<|/ref|><|det|>[[116, 153, 880, 201]]<|/det|>
+Maybe, the authors could add the results of comparative statistics tests on barplots and boxplots (either directly report the p- values or show corresponding labels as \*\*\* or \*\*\*)?  
+
+<|ref|>text<|/ref|><|det|>[[116, 202, 880, 317]]<|/det|>
+Do they have any hypothesis to explain the much higher concentration in reactive C(VI) measured in the A7 surface soil- ash sample from the serpentine chaparral landscape ? The data provided in Table S6 indicate about twice more total Cr in this sample compared to sample A6 (for instance), but the concentration in reactive Cr(VI) is more than 3 times higher in the bulk fraction and more than 25 three times higher in the \(< 53\mu m\) fraction. Are there any mineralogical differences with the other moderatelyhighly burned serpentine soil- ash samples (A3- A6) that could help to explain that ?  
+
+<|ref|>sub_title<|/ref|><|det|>[[117, 334, 194, 350]]<|/det|>
+## Figure 4  
+
+<|ref|>text<|/ref|><|det|>[[116, 353, 880, 435]]<|/det|>
+I wonder if this figure should be maintained in the main text. First, I am not convince that it really supports the assumption that 'total Cr(VI) was most abundant in wind- dispersible soils and ash particulates after high fire severity conditions compared to a low severity sample'. Second, I do not understand why the two figures are plotted on different sizes (A is larger than B).  
+
+<|ref|>text<|/ref|><|det|>[[116, 435, 880, 484]]<|/det|>
+It could maybe be replaced by a figure similar to Figure S7 (maybe use only panels A and B or panels C and D from Figure S7 and add two similar panels from a low fire severity burn sample from a serpentine chaparral soil)?  
+
+<|ref|>text<|/ref|><|det|>[[116, 484, 880, 581]]<|/det|>
+Whatever, if Figure 4 is maintained in the main text, I would recommend that the authors add some histograms with the estimated number of Cr(VI) and total Cr particles on the two figures, in order to help the reader to better assess the relative proportion of both types of particles. At least, I would recommend that they change for more contrasted colors, in order to help the reader to better visually decipher between Cr(VI) and total Cr particles.  
+
+<|ref|>sub_title<|/ref|><|det|>[[117, 598, 208, 614]]<|/det|>
+## Figure S6  
+
+<|ref|>text<|/ref|><|det|>[[116, 617, 880, 650]]<|/det|>
+I would recommend to change the letters for A, B and C in the legend to fit with the letters reported on the figure.  
+
+<|ref|>sub_title<|/ref|><|det|>[[117, 667, 201, 682]]<|/det|>
+## Table S6  
+
+<|ref|>text<|/ref|><|det|>[[116, 686, 880, 899]]<|/det|>
+Total concentrations in bulk surface soil- ash samples (and some \(< 53\mu m\) fractions) are displayed, but total concentration in bulk soil samples across the whole soil cores are not provided. I would recommend that the authors provide these data (at least those for Cr) as mean total concentration for rhyolite, melange and serpentine soils in SI, either in the form of a Table or as a figure similar to Figure 2. Such a figure would better show that the fractions of reactive Cr(VI) is very low compared to total Cr concentration. In the same way, I would have been interested to see a figure similar to Figure 2 that would have depicted the fraction of reactive Cr(VI) as a function of the total Cr concentration. Indeed, such a figure would have helped to check if the fraction of reactive Cr(VI) is really higher in serpentine soils. Even if I agree with the authors that the concentration of reactive Cr(VI) is the most relevant parameter to assess a potential environmental and/or health risk, the fraction of reactive Cr(VI) could further inform on the actual mechanism(s) and/or soil characteristic(s) that favor Cr(III) to
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 82, 880, 117]]<|/det|>
+Cr(VI) oxidation in burned soils. But, maybe this question is beyond the scope of the paper...  
+
+<|ref|>sub_title<|/ref|><|det|>[[117, 132, 208, 149]]<|/det|>
+## Figure S7  
+
+<|ref|>text<|/ref|><|det|>[[117, 152, 880, 234]]<|/det|>
+Figure S7Legend : ... from a serpentine chaparral soil...Why did the authors not tried to analyze more some Cr(VI) areas on panel A ?The XANES spectrum at point 1 on panel C shows a well-marked Cr(VI) pre-edge peak but the color code indicate rather low amounts of Cr(VI) at this point. Could the authors explain that ?  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 250, 410, 267]]<|/det|>
+## Suggested additional references  
+
+<|ref|>text<|/ref|><|det|>[[115, 270, 883, 530]]<|/det|>
+Suggested additional referencesRascio et al., 2022. Evidence of hexavalent chromium formation and changes of Cr speciation after laboratory- simulated fires of composted tannery sludges long- term amended agricultural soils. Journal of Hazardous Materials, 436, 129117. https://doi.org/10.1016/j.jhazmat.2022.129117Terzano et al., 2021. Fire effects on the distribution and bioavailability of potentially toxic elements (PTEs) in agricultural soils. Chemosphere, 130752. https://doi.org/10.1016/j.chemosphere.2021.130752Ré et al., 2021. Cytotoxic effects of wildfires ashes : In- vitro responses of skin cells. Environmental Pollution, 285, 117279. https://doi.org/10.1016/j.envpol.2021.117279Jahn et al., 2021. Metallic and crustal elements in biomass- burning aerosols and ash: Prevalence, significance, and similarities to soil particles. ACS Earth and Space Chemistry, 5, 136- 148. https://dx.doi.org/10.1021/acsearthspacechem.0c00191Xu et al., 2020. Wildfires, global climate change, and human health. The New England Journal of Medicine, 383, 2173- 2181. https://doi.org/10.1056/NEJMsr2028985
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[119, 85, 437, 101]]<|/det|>
+## Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[117, 103, 880, 252]]<|/det|>
+The authors have to highlight the novelty of their manuscript. The abstract should be revised to attract the reader's attention. The Introduction section should be improved by adding references dealing soil contamination issues. However, the problem is that the English and the whole organisation of the present version are definitely below an acceptable standard for an international scientific journal. Analytical quality control is missing. Detection limits of the applied methods should be reported. The main problem for this manuscript is its structure. Major parts are missing from the manuscript. My suggestion is to reject this manuscript and encourage the authors to submit a more mature manuscript.
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[345, 91, 651, 111]]<|/det|>
+## Response to Review Comments  
+
+<|ref|>sub_title<|/ref|><|det|>[[115, 128, 220, 146]]<|/det|>
+## Reviewer #1  
+
+<|ref|>text<|/ref|><|det|>[[115, 162, 880, 355]]<|/det|>
+This manuscript presents the results of a study aimed at emphasizing the health threat potentially arising from chromium transformation to its most harmful \(\mathrm{Cr(VI)}\) form in soils and ashes upon wildfires. To reach their objective, the authors analyzed 38 cores drilled in more or less severely burned soils at four Preserves that experienced large wildfires in the North Coast Range of California (USA). The results obtained confirm the catalytic effect of high temperatures generated by wildfires on chromium oxidation in soil, and they point to control of soil geology and fire severity on this effect. They also interestingly indicate that reactive \(\mathrm{Cr(VI)}\) can reach dangerous levels in wind- dispersible particulates found in the surficial layers of the soil at ultramafic settings, and that this harmful form of chromium can persist in the soil/ash system up to one year after the wildfire if rainfalls are not significant. These results yield to the conclusion that more work is required to further evaluate this potential risk around the world.  
+
+<|ref|>text<|/ref|><|det|>[[115, 371, 880, 477]]<|/det|>
+I found this paper very pleasant to read. The subject is well introduced, the sites and samples are well described and fit with the objective of the study, the results are well presented and they support the discussion that gives proper attention to the existing literature on the topic and yield to the novel conclusions that geology and fire intensity are important drivers of harmful hexavalent chromium in soils and wind- dispersible soil/ash surface particles, and that the associated health risk can persist up to one year after fire.  
+
+<|ref|>text<|/ref|><|det|>[[115, 493, 881, 633]]<|/det|>
+Although the scale of the study (North Coast Range of California, USA) could at first be considered too short to extend its findings at the global scale, I agree with the authors that the number and diversity of sites studied can be considered enough to address this issue. I also agree with the concluding remark on the importance of further evaluating the potential risk of wildfire- induced harmful hexavalent chromium in wind- dispersible soil/ash particles at the global scale. Finally, I agree with the authors that the results provided in this study deserve to be shared with a large audience, ranging from scientists working on the topic to policy makers, public administrators and nature managers.  
+
+<|ref|>text<|/ref|><|det|>[[116, 650, 837, 703]]<|/det|>
+For all these reasons, I consider that this manuscript deserves to be published in Nature Communications. I have listed below few issues that I would be interested to see addressed, although only few of them are mandatory for publication.  
+
+<|ref|>text<|/ref|><|det|>[[173, 719, 872, 771]]<|/det|>
+Response: We thank the reviewer for their constructive feedback and review of the manuscript's findings and implications. We have reviewed and addressed in detail below the reviewer's suggestions.  
+
+<|ref|>text<|/ref|><|det|>[[171, 788, 790, 807]]<|/det|>
+Changes: Please see our specific changes to the reviewer's suggestions below.  
+
+<|ref|>text<|/ref|><|det|>[[115, 823, 879, 894]]<|/det|>
+Figure 1: I agree that a large fraction of ultramafic/mafic areas are concerned by wildfires at the global scale. Figure 1 shows that these areas are mainly located in the tropical region where most soils are deeply weathered (Ultisols and Oxisols, according to the USDA classification). However, the Fe contents reported in Table S6 suggest that the soils studied in this work do not
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 90, 850, 143]]<|/det|>
+correspond to these soil types. This raises the question of the actual representativeness of the results regarding tropical areas. I would recommend that the authors comment on that point, and maybe further consider it in their concluding remarks.  
+
+<|ref|>text<|/ref|><|det|>[[172, 160, 880, 440]]<|/det|>
+Response: We agree with the reviewer that the soils in this study are not as deeply weathered as Oxisols and Ultisols, like lateritic soils, common across tropical climate regions (e.g., New Caledonia, Cuba, Brazil, Malaysia, Indonesia, Madagascar, northern Australia). In our study, the average Fe concentration across burned and unburned bulk ultramafic (serpentine) soil was 8.71 wt. \(\%\) , while lateritic soils can contain more than 5x more Fe, typically in the form of crystalline Fe oxides (e.g., hematite, goethite). Chromium is predominantly found within these crystalline Fe oxides. Additionally, lateritic soils are often depleted in Ca, Mg, and Si. The Fe concentrations of our study soils are representative of serpentine soils in Mediterranean or temperate climates (e.g., California, Oregon, Washington, Turkey, Balkans) with moderate Fe oxide content and neutral to alkaline pH. We have added a paragraph within the Discussion section that addresses the characterization of ultramafic soils globally and the limitations of soil types within our study site when considering mechanisms in lateritic and highly weathered metal- rich soils in the tropical regions. Furthermore, we place our findings from field analysis in the context of laboratory studies that examine Cr(VI) generation from Cr(III) solids common to lateritic soils.  
+
+<|ref|>text<|/ref|><|det|>[[173, 456, 546, 473]]<|/det|>
+Changes: Within the Discussion we now state:  
+
+<|ref|>text<|/ref|><|det|>[[173, 473, 877, 769]]<|/det|>
+"While recognized for urban fires, threats from metal exposure in smoke and dust need to be recognized within wildland fires arising on metal- rich geologies. Across tropical climate regions, deeply weathered lateritic soils are common, in which Cr is predominantly found within crystalline Fe oxides (e.g., hematite, goethite), and soil Fe content may exceed 50 wt. \(\%\) . Past work has quantified the Cr oxidation capacity of Cr- bearing Fe oxides and lateritic soils during heating simulations \(^{35,62}\) . For example, the greatest Cr(VI) formation (more than 40% of total Cr) upon heating hematite occurred at temperatures less than \(400^{\circ}\mathrm{C}\) , while up to 100% of total Cr in goethite transformed to Cr(VI) at \(800^{\circ}\mathrm{C}\) . In Mediterranean or temperate climates (similar to this study's region), ultramafic soils are relatively more enriched in Cr- bearing phyllosilicate minerals (e.g., serpentine), Fe oxide content is moderate (typically less than 10 wt. \(\%\) ) with more amorphous secondary phases, and soil pH is neutral to alkaline pH. In our study, and under natural wildfire conditions, we observed that up to about 0.015% of total Cr was reactive Cr(VI) in burned serpentine soils. Chromium(VI) generation during wildfires depend on fire conditions and host mineralogy; thus, the extent of Cr(VI) formation in lateritic soils may differ from temperate serpentine soils, but field observations of Cr(VI) in burned lateritic soils following wildfires are currently lacking."  
+
+<|ref|>text<|/ref|><|det|>[[115, 785, 857, 821]]<|/det|>
+Figure 3: Maybe, the authors could add the results of comparative statistics tests on barplots and boxplots (either directly report the \(p\) - values or show corresponding labels as \(***\) or \(**\) )?  
+
+<|ref|>text<|/ref|><|det|>[[174, 838, 856, 890]]<|/det|>
+Response: We performed a one- way ANOVA for Figure 3c and found no statistically significant difference between the three groups (p- value = 0.164). In Figure 3b, we are illustrate reactive concentrations within surface soil- ash based on fire severity (low
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[174, 90, 877, 143]]<|/det|>
+versus moderate- high) and particle size (bulk soil less than \(2\mathrm{mm}\) versus the silt and clay- sized fraction less than \(53\mu \mathrm{m}\) ). We are unable to run a two- way ANOVA because burn severity sample sizes are not equal.  
+
+<|ref|>text<|/ref|><|det|>[[173, 160, 879, 247]]<|/det|>
+Changes: For Figure 3, we now state "Reactive \(\mathrm{Cr(VI)}\) concentrations ranged from 64 to \(1,060\mu \mathrm{g / kg}\) , with a median concentration of \(257\mu \mathrm{g / kg}\) , and remained elevated compared to concentrations (5- 64 \(\mu \mathrm{g / kg}\) ) within the near surface depths (0- 2 cm) of unburned serpentine soil (Figure 3c); however, these differences were not statistically significant (p- value \(= 0.164\) )."  
+
+<|ref|>text<|/ref|><|det|>[[113, 263, 880, 387]]<|/det|>
+Do they have any hypothesis to explain the much higher concentration in reactive \(\mathrm{C(VI)}\) measured in the A7 surface soil- ash sample from the serpentine chaparral landscape? The data provided in Table S6 indicate about twice more total \(\mathrm{Cr}\) in this sample compared to sample A6 (for instance), but the concentration in reactive \(\mathrm{Cr(VI)}\) is more than 3 times higher in the bulk fraction and more than 25 three times higher in the \(< 53\mu \mathrm{m}\) fraction. Are there any mineralogical differences with the other moderately- highly burned serpentine soil- ash samples (A3- A6) that could help to explain that?  
+
+<|ref|>text<|/ref|><|det|>[[173, 403, 877, 508]]<|/det|>
+Response: At site A7, the soil experienced longer burning duration and fire intensities with greater biomass combustion that may further contribute to the high- levels of \(\mathrm{Cr(VI)}\) . Unlike A7 (original Figure S9; revised Figure S11), we did not observe mineralogical changes in bulk composition for samples A3- A6 in the surface soil and ash compared to underlying burned soil. We suspect that high temperatures did not persist for sufficient time to alter bulk mineralogy in the latter samples.  
+
+<|ref|>text<|/ref|><|det|>[[173, 525, 877, 682]]<|/det|>
+Changes: Within the Results we state: "At site A7, the average reactive \(\mathrm{Cr(VI)}\) concentration was more than three times greater than other moderate- high fire severity sites (Figure 3b). We suspect that longer burning duration and fire intensities with greater biomass combustion contributed to the relatively high- levels of \(\mathrm{Cr(VI)}\) , as this was a severely burned forested area. Importantly, ash from severely burned areas concentrate alkali (Na, K) and alkaline earth (Ca, Mg) metals (often from biomass combustion) that are key for the thermal oxidation of \(\mathrm{Cr(III)}\) \(^{13,43}\) . For example, \(\mathrm{CaCrO_4}\) was noted after agricultural soil amended with composted \(\mathrm{Cr(III)}\) - rich tannery sludge was heated at \(500^{\circ}\mathrm{C}\) \(^{55}\) ."  
+
+<|ref|>text<|/ref|><|det|>[[114, 698, 880, 787]]<|/det|>
+Figure 4: I wonder if this figure should be maintained in the main text. First, I am not convince that it really supports the assumption that "total \(\mathrm{Cr(VI)}\) was most abundant in wind- dispersible soils and ash particulates after high fire severity conditions compared to a low severity sample". Second, I do not understand why the two figures are plotted on different sizes (A is larger than B).  
+
+<|ref|>text<|/ref|><|det|>[[113, 803, 875, 857]]<|/det|>
+It could maybe be replaced by a figure similar to Figure S7 (maybe use only panels A and B or panels C and D from Figure S7 and add two similar panels from a low fire severity burn sample from a serpentine chaparral soil)?
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 90, 865, 178]]<|/det|>
+Whatever, if Figure 4 is maintained in the main text, I would recommend that the authors add some histograms with the estimated number of Cr(VI) and total Cr particles on the two figures, in order to help the reader to better assess the relative proportion of both types of particles. At least, I would recommend that they change for more contrasted colors, in order to help the reader to better visually decipher between Cr(VI) and total Cr particles.  
+
+<|ref|>text<|/ref|><|det|>[[173, 194, 870, 334]]<|/det|>
+Response: We appreciate the reviewer's feedback regarding Figure 4. We agree that our results of greater Cr(VI)- bearing particles in high severity soil- ash versus low- severity conditions can be clearer, especially by including a histogram of particles containing Cr(VI). Regarding the reviewer's concern about the different size plots, the sample area for large- scale micro- XRF maps of each thin section were not held constant during data collection, resulting in different mapped areas. Due to limitations of the analysis software, we are unable to change for more contrasted colors, but we can update the min/max values for color brightness.  
+
+<|ref|>text<|/ref|><|det|>[[173, 350, 870, 473]]<|/det|>
+We have revised Figure 4 to focus on a \(1 - \mu \mathrm{m}\) resolution XRF image containing particles from a high fire severity sample with and without measurable Cr(VI) by XANES analysis. We then moved the original Figure 4 to the Supplementary Information as Figure S8, and added another example of Cr(VI)- containing particles in a \(1 - \mu \mathrm{m}\) resolution XRF image as Figure S9. We also revised the main text related to the figures to highlight the presence of Cr(VI)- containing particles within the high fire severity samples, which was not apparent in low fire severity samples.  
+
+<|ref|>text<|/ref|><|det|>[[174, 490, 682, 508]]<|/det|>
+Changes: See revised Figure 4, Figure S8, and Figure S9 below.  
+
+<|ref|>text<|/ref|><|det|>[[173, 524, 872, 630]]<|/det|>
+"Using micro- scale X- ray techniques, Cr(VI)- containing soil and ash particulates were identified in a high fire severity sample as opposed to a low fire severity sample (Figure 4; Figure S8). Here, Cr(VI) was associated with mineral surfaces (e.g., adsorbed) or enriched in relatively low- Cr particles with Ca and K (Figure 4, Figure S9). Consistent with particle analysis, reactive Cr(VI) concentrations spanned from 326 to \(13,000 \mu \mathrm{g / kg}\) (Figure 3b)."
+
+<--- Page Split --->
+<|ref|>image<|/ref|><|det|>[[216, 156, 753, 740]]<|/det|>
+<|ref|>image_caption<|/ref|><|det|>[[125, 750, 864, 891]]<|/det|>
+<center>Figure 4 | Total Cr(VI) in wind-dispersible soil and ash particles. a. \(\mu\) -XRF image (pixel resolution: \(1 \mu m\) ) showing the relative intensity of Cr(VI) (green; estimated as the intensity ratio at 5993 and 6010 eV) and total Cr (blue; measured at 6010 eV) within the \(< 53 - \mu m\) size fraction of Cr-bearing soil-ash particulates from a serpentine chaparral that experienced high fire severity (A7). b. Normalized \(\mu\) -XANES spectra (Cr K-edge) from numbered locations on Cr-bearing particles in a. Dashed lines indicate energies characteristic of Cr(VI) (5993 eV), Cr(III) (6003 eV), and total Cr (6010 eV), at which \(\mu\) -XRF images were also collected. </center>
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[113, 90, 875, 179]]<|/det|>
+Figure S8 | Total Cr(VI) in wind-dispersible soil and ash particles. \(\mu\) - XRF image showing particle distribution of total Cr(VI) (green; estimated as the intensity ratio at 5993 and 6010 eV) and total Cr (blue; measured at 6010 eV) within the \(< 53 - \mu \mathrm{m}\) size fraction of Cr-bearing soil- ash particulates from a. high fire severity site (A7) and b. low fire severity site (A1) in a serpentine chaparral.  
+
+<|ref|>image<|/ref|><|det|>[[128, 182, 777, 444]]<|/det|>
+<|ref|>image_caption<|/ref|><|det|>[[113, 466, 878, 590]]<|/det|>
+<center>Figure S9 | Total Cr(VI) in wind-dispersible soil and ash particles. a. \(\mu\) -XRF image (pixel resolution: \(1\mu \mathrm{m}\) ) showing the relative intensity of \(\mathrm{Cr(VI)}\) (green; estimated as the intensity ratio at 5993 and \(6010\mathrm{eV}\) ) and total Cr (blue; measured at \(6010\mathrm{eV}\) ) within the \(< 53 - \mu \mathrm{m}\) size fraction of Cr-bearing soil-ash particulates from a serpentine chaparral that experienced high fire severity (A7). b. Normalized \(\mu\) -XANES spectra (Cr K-edge) from numbered locations on Cr-bearing particles in a. Dashed lines indicate energies characteristic of \(\mathrm{Cr(VI)}\) (5993 eV), \(\mathrm{Cr(III)}\) (6003 eV), and total Cr (6010 eV). </center>  
+
+<|ref|>image<|/ref|><|det|>[[128, 599, 644, 875]]<|/det|>
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 90, 866, 125]]<|/det|>
+Figure S6: I would recommend to change the letters for A, B and C in the legend to fit with the letters reported on the figure.  
+
+<|ref|>text<|/ref|><|det|>[[173, 142, 864, 194]]<|/det|>
+Response: Based on the reviewer's recommendation, we have revised the figure, accordingly, by changing uppercase letters to lowercase, in addition to the other figures with sub- panels (similarly identified by Reviewer #2).  
+
+<|ref|>text<|/ref|><|det|>[[173, 212, 652, 229]]<|/det|>
+Changes: We changed the letters to lowercase, as requested.  
+
+<|ref|>text<|/ref|><|det|>[[115, 246, 877, 352]]<|/det|>
+Table S6: Total concentrations in bulk surface soil- ash samples (and some \(< 53 \mu \mathrm{m}\) fractions) are displayed, but total concentration in bulk soil samples across the whole soil cores are not provided. I would recommend that the authors provide these data (at least those for Cr) as mean total concentration for rhyolite, mélange and serpentine soils in SI, either in the form of a Table or as a figure similar to Figure 2. Such a figure would better show that the fractions of reactive \(\mathrm{Cr(VI)}\) is very low compared to total Cr concentration.  
+
+<|ref|>text<|/ref|><|det|>[[173, 368, 880, 437]]<|/det|>
+Response: We thank the reviewer for their recommendation, which we have addressed in the revision. Please refer to our response and changes to the reviewer's next comment related to including a figure showing the fraction of reactive \(\mathrm{Cr(VI)}\) to total Cr concentrations.  
+
+<|ref|>text<|/ref|><|det|>[[173, 455, 881, 542]]<|/det|>
+Changes: Using total Cr concentrations reported for soil cores in Table S1, we have revised Table S6 to include the mean total element concentrations for Cr, Fe, Mn, Ni, Ca, Mg, Na, and K in addition to the surface soil- ash sample data so that the reader can compare elemental concentrations in surface soil and ash to bulk soil from different geologies.
+
+<--- Page Split --->
+<|ref|>table_caption<|/ref|><|det|>[[114, 91, 875, 160]]<|/det|>
+Table S6 | Physicochemical characteristics of bulk soil and ash (up to 2 mm), and selected fine size fractions less than 53 μm, collected from surface layers of the burned serpentine chaparral, and mean elemental concentrations from bulk underlying soil based on geology type (rhyolitic, mélange, and serpentine).
+
+<|ref|>table<|/ref|><|det|>[[117, 193, 892, 530]]<|/det|>
+
+<table><tr><td>ID</td><td>Fire<br>Severitya</td><td>% Sandb<br>(2-0.05 mm)</td><td>% Siltc<br>(53-2 μm)</td><td>% Clayc<br)&lt;2 μm)</td><td>Cr<br>mg/kg</td><td>Fe<br>mg/g</td><td>Mn<br>mg/kg</td><td>Ni<br>mg/kg</td><td>Ca<br>mg/g</td><td>Mg<br>mg/g</td><td>Na<br>mg/g</td><td>K<br>mg/g</td></tr><tr><td>Surface Soil-Ash</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td>A1</td><td>L</td><td></td><td></td><td></td><td>1147</td><td>64.7</td><td>1351</td><td>1528</td><td>15.5</td><td>77.3</td><td>4.63</td><td>3.76</td></tr><tr><td>A2</td><td>L</td><td></td><td></td><td></td><td>1532</td><td>69.1</td><td>1203</td><td>2530</td><td>10.2</td><td>158</td><td>&lt;0.1</td><td>1.65</td></tr><tr><td>A3</td><td>M/H</td><td></td><td></td><td></td><td>1606</td><td>84.2</td><td>1480</td><td>3117</td><td>9.3</td><td>148</td><td>&lt;0.1</td><td>1.54</td></tr><tr><td>A4</td><td>M/H</td><td>87.5</td><td>11.7</td><td>0.8</td><td>2256</td><td>78.9</td><td>1438</td><td>3380</td><td>8.7</td><td>159</td><td>&lt;0.1</td><td>0.90</td></tr><tr><td>A5</td><td>M/H</td><td></td><td></td><td></td><td>999</td><td>57.7</td><td>1022</td><td>1849</td><td>10.2</td><td>125</td><td>2.10</td><td>2.48</td></tr><tr><td>A6</td><td>M/H</td><td>78.2</td><td>20.2</td><td>1.6</td><td>1970</td><td>79.2</td><td>1555</td><td>2726</td><td>29.9</td><td>174</td><td>&lt;0.1</td><td>2.86</td></tr><tr><td>A7</td><td>M/H</td><td>85.7</td><td>12.9</td><td>1.4</td><td>4829</td><td>102</td><td>1543</td><td>2643</td><td>17.9</td><td>211</td><td>&lt;0.1</td><td>2.29</td></tr><tr><td>Less than 53 μm size fraction</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td>A4</td><td>M/H</td><td></td><td></td><td></td><td>1133</td><td>94.7</td><td>1975</td><td>3042</td><td>41.5</td><td>124</td><td>&lt;0.1</td><td>3.38</td></tr><tr><td>A6</td><td>M/H</td><td></td><td></td><td></td><td>946</td><td>82.7</td><td>1910</td><td>2294</td><td>44.2</td><td>133</td><td>&lt;0.1</td><td>4.41</td></tr><tr><td>A7</td><td>M/H</td><td></td><td></td><td></td><td>1643</td><td>105</td><td>2547</td><td>3489</td><td>34.0</td><td>189</td><td>2.02</td><td>3.91</td></tr><tr><td>Bulk Soild</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td>Rhyolitic \((n=7)\)</td><td></td><td></td><td></td><td></td><td>162</td><td>34.5</td><td>796</td><td>88</td><td>10.7</td><td>7.9</td><td>13.6</td><td>9.70</td></tr><tr><td>Melange \((n=16)\)</td><td></td><td></td><td></td><td></td><td>314</td><td>50.0</td><td>855</td><td>259</td><td>10.7</td><td>37.3</td><td>8.17</td><td>13.2</td></tr><tr><td>Serpentine \((n=10)\)</td><td></td><td></td><td></td><td></td><td>2373</td><td>87.1</td><td>1511</td><td>2929</td><td>4.4</td><td>150</td><td>3.61</td><td>1.69</td></tr></table>
+
+<|ref|>text<|/ref|><|det|>[[114, 537, 511, 550]]<|/det|>
+a L = Low severity, M = Moderate severity, H = High severity
+
+<|ref|>text<|/ref|><|det|>[[114, 552, 325, 564]]<|/det|>
+b Determined by sieve analysis
+
+<|ref|>text<|/ref|><|det|>[[114, 567, 464, 580]]<|/det|>
+c Determined by laser diffraction particle size counter
+
+<|ref|>text<|/ref|><|det|>[[114, 580, 854, 606]]<|/det|>
+d Bulk soil concentrations are mean values using all soil cores (fire-affected and unburned) for each geology type:rhyolitic \((n=7)\) , mélange \((n=16)\) , and serpentine \((n=10)\) .
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[113, 89, 878, 232]]<|/det|>
+In the same way, I would have been interested to see a figure similar to Figure 2 that would have depicted the fraction of reactive \(\mathrm{Cr(VI)}\) as a function of the total \(\mathrm{Cr}\) concentration. Indeed, such a figure would have helped to check if the fraction of reactive \(\mathrm{Cr(VI)}\) is really higher in serpentine soils. Even if I agree with the authors that the concentration of reactive \(\mathrm{Cr(VI)}\) is the most relevant parameter to assess a potential environmental and/or health risk, the fraction of reactive \(\mathrm{Cr(VI)}\) could further inform on the actual mechanism(s) and/or soil characteristic(s) that favor \(\mathrm{Cr(III)}\) to \(\mathrm{Cr(VI)}\) oxidation in burned soils. But, maybe this question is beyond the scope of the paper...  
+
+<|ref|>text<|/ref|><|det|>[[172, 247, 880, 456]]<|/det|>
+Response: We thank the reviewer for their recommendation. We agree that the reactive \(\mathrm{Cr(VI)}\) fraction of total \(\mathrm{Cr}\) in soil and soil- ash is low; however, the fraction is relatively higher in near surface soil depths compared to past studies quantifying natural \(\mathrm{Cr(III)}\) oxidation in unburned soils, including a 2017 study at McLaughlin Natural Reserve. Moreover, as the reviewer notes, the hazard imposed by the particulates is related to the reactive \(\mathrm{Cr(VI)}\) . The percentage of total \(\mathrm{Cr}\) that was reactive \(\mathrm{Cr(VI)}\) in unburned serpentine soil was consistent with previous measurements (McClain et al., 2017). Interestingly, the reactive \(\mathrm{Cr(VI)}\) fraction differs based on geology (rhyolite, mélange, and serpentinite). The reactive \(\mathrm{Cr(VI)}\) fraction in rhyolitic and mélange soils composed more of the total \(\mathrm{Cr}\) content than the relative fraction within serpentine soils. In order to highlight these variations across soil depth in fire- affected and unburned sites, we added a figure to the supplementary information.  
+
+<|ref|>text<|/ref|><|det|>[[174, 473, 866, 526]]<|/det|>
+Changes: Figure S6 (below) was added to the Supplementary Information. Succeeding figure numbers were updated based on this addition. Within the Results and Discussion sections, we now state:  
+
+<|ref|>text<|/ref|><|det|>[[174, 542, 877, 629]]<|/det|>
+"Average \(\mathrm{Cr(VI)}\) concentrations generated in soils derived from mélange (Figure 2b) were more than double the respective levels in rhyolitic soil (Figure 2a) regardless of overlapping ranges in total \(\mathrm{Cr}\) content, 152- 954 and 102- 338 mg/kg, respectively, reflecting the potential contribution of differing mineralogy to \(\mathrm{Cr(VI)}\) generation (Figure S6) 35. "  
+
+<|ref|>text<|/ref|><|det|>[[172, 647, 881, 682]]<|/det|>
+"In our study, and under natural wildfire conditions, we observed that up to about \(0.015\%\) of total \(\mathrm{Cr}\) was reactive \(\mathrm{Cr(VI)}\) in burned serpentine soils (Figure S6)."
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[113, 90, 875, 211]]<|/det|>
+Figure S6 | Fraction of total Cr that is reactive \(\mathrm{Cr(VI)}\) in burned and unburned soils. The ratio of reactive \(\mathrm{Cr(VI)}\) to total Cr concentrations (as a percentage) within a. rhyolite- , b. mélange- , c. serpentinite- derived soil profiles (0- 16 cm) that were not burned (gray; rhyolite, \(n =\) 3; mélange, \(n = 7\) ; serpentinite, \(n = 3\) ) or were fire- affected (colored; rhyolite, \(n = 4\) ; mélange, \(n\) \(= 9\) ; serpentinite, \(n = 7\) ). Percentages were also plotted for serpentinite- derived soil (0- 20 cm) from McClain et al. (2017) in c. for comparison. Each point represents the average percentage for a soil core based on triplicate measurements.  
+
+<|ref|>image<|/ref|><|det|>[[125, 222, 870, 461]]<|/det|>
+<|ref|>image_caption<|/ref|><|det|>[[115, 480, 572, 500]]<|/det|>
+<center>Figure S7: Legend : ... from a serpentine chaparral soil... </center>  
+
+<|ref|>text<|/ref|><|det|>[[172, 515, 880, 569]]<|/det|>
+Response: We agree with the reviewer's recommendation and have made the appropriate changes. Based on the reviewer's previous recommendations, this figure has revised as Figure 4 and Figure S9.  
+
+<|ref|>text<|/ref|><|det|>[[172, 585, 852, 621]]<|/det|>
+Changes: (Figure 4 and Figure S9 caption) "... particles ( \(< 53 \mu \mathrm{m}\) ) from a serpentine chaparral that experienced high fire severity ..."  
+
+<|ref|>text<|/ref|><|det|>[[115, 638, 870, 691]]<|/det|>
+Why did the authors not tried to analyze more some \(\mathrm{Cr(VI)}\) areas on panel A? The XANES spectrum at point 1 on panel C shows a well- marked \(\mathrm{Cr(VI)}\) pre- edge peak but the color code indicate rather low amounts of \(\mathrm{Cr(VI)}\) at this point. Could the authors explain that?  
+
+<|ref|>text<|/ref|><|det|>[[172, 708, 875, 760]]<|/det|>
+Response: Our XANES analysis was used to corroborate our bulk measurements of reactive \(\mathrm{Cr(VI)}\) . Particles within the XRF map were used to denote the presence and abundance of \(\mathrm{Cr(VI)}\) , corroborating (and visualizing) the reactive fraction measurements.  
+
+<|ref|>text<|/ref|><|det|>[[172, 777, 857, 830]]<|/det|>
+Changes: "To corroborate bulk measurements of reactive \(\mathrm{Cr(VI)}\) , we combined multi- energy mapping with \(\mu\) - XANES of select spots on particles to confirm the presence of \(\mathrm{Cr(VI)}\) ."  
+
+<|ref|>sub_title<|/ref|><|det|>[[115, 848, 375, 865]]<|/det|>
+## Suggested additional references  
+
+<|ref|>text<|/ref|><|det|>[[115, 866, 861, 901]]<|/det|>
+Rascio et al., 2022. Evidence of hexavalent chromium formation and changes of \(\mathrm{Cr}\) speciation after laboratory- simulated fires of composted tannery sludges long- term amended agricultural
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[113, 88, 868, 300]]<|/det|>
+soils. Journal of Hazardous Materials, 436, 129117. https://doi.org/10.1016/j.jhazmat.2022.129117 Terzano et al., 2021. Fire effects on the distribution and bioavailability of potentially toxic elements (PTEs) in agricultural soils. Chemosphere, 130752. https://doi.org/10.1016/j.chemosphere.2021.130752 Ré et al., 2021. Cytotoxic effects of wildfires ashes : In- vitro responses of skin cells. Environmental Pollution, 285, 117279. https://doi.org/10.1016/j.envpol.2021.117279 Jahn et al., 2021. Metallic and crustal elements in biomass- burning aerosols and ash: Prevalence, significance, and similarities to soil particles. ACS Earth and Space Chemistry, 5, 136- 148. https://dx.doi.org/10.1021/acsearthspacechem.0c00191 Xu et al., 2020. Wildfires, global climate change, and human health. The New England Journal of Medicine, 383, 2173- 2181. https://doi.org/10.1056/NEJMsr2028985  
+
+<|ref|>text<|/ref|><|det|>[[172, 315, 868, 352]]<|/det|>
+Response: We thank the reviewer for sharing additional references. We agree that these references are relevant to the study.  
+
+<|ref|>text<|/ref|><|det|>[[173, 368, 857, 404]]<|/det|>
+Changes: We have added these references suggested above, in addition to a few other recently published and relevant studies, to discussions within the main text.  
+
+<|ref|>text<|/ref|><|det|>[[173, 420, 866, 455]]<|/det|>
+"Global wildfire activity represents a rising distributed health risk from smoke and dust inhalation \(^{5 - 10}\) ."  
+
+<|ref|>text<|/ref|><|det|>[[173, 472, 870, 508]]<|/det|>
+10. Xu, R. et al. Wildfires, Global Climate Change, and Human Health. N. Engl. J. Med. 383, 2173-2181 (2020).  
+
+<|ref|>text<|/ref|><|det|>[[173, 524, 872, 575]]<|/det|>
+"Increased heavy metals in PM have been documented during wildfire episodes and may induce cytotoxicity, increase lung cancer risks, and greatly contribute to oxidative stress \(^{19 - 30}\) ."  
+
+<|ref|>text<|/ref|><|det|>[[174, 593, 830, 648]]<|/det|>
+"Suburban fires illustrates the impacts of inhaling Cr(VI)- containing ash within the respiratory tract by measuring Cr(VI) leached with a simulated lung fluid \(^{36,37}\) and discerning Cr mineralogy within nano- sized particulates (< 100 nm) \(^{22,30}\) ."  
+
+<|ref|>text<|/ref|><|det|>[[174, 664, 861, 717]]<|/det|>
+21. Boaggio, K. et al. Beyond Particulate Matter Mass: Heightened Levels of Lead and Other Pollutants Associated with Destructive Fire Events in California. Environ. Sci. Technol. 56, 14272-14283 (2022).  
+
+<|ref|>text<|/ref|><|det|>[[173, 733, 870, 769]]<|/det|>
+22. Alshehri, T. et al. Wildland-urban interface fire ashes as a major source of incidental nanomaterials. J. Hazard. Mater. 443, 130311 (2023).  
+
+<|ref|>text<|/ref|><|det|>[[174, 785, 860, 838]]<|/det|>
+28. Jahn, L. G. et al. Metallic and crustal elements in biomass-burning aerosol and ash: Prevalence, significance, and similarity to soil particles. ACS Earth Sp. Chem. 5, 136-148 (2021).  
+
+<|ref|>text<|/ref|><|det|>[[173, 855, 830, 890]]<|/det|>
+29. Ré, A. et al. Cytotoxic effects of wildfire ashes: In-vitro responses of skin cells. Environ. Pollut. 285, 117279 (2021).
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[174, 106, 880, 160]]<|/det|>
+"Metals in soils and ash are commonly linked to structural burning within wildland- urban interfaces (WUI) \(^{1,31 - 33}\) , with negligible awareness of wildland landscapes (soils and ash) as an alternative and highly distributed source \(^{1}\) ."  
+
+<|ref|>text<|/ref|><|det|>[[174, 177, 880, 230]]<|/det|>
+32. Alexakis, D. E. Suburban areas in flames: Dispersion of potentially toxic elements from burned vegetation and buildings. Estimation of the associated ecological and human health risk. Environ. Res. 183, 109153 (2020).  
+
+<|ref|>text<|/ref|><|det|>[[174, 246, 867, 299]]<|/det|>
+33. Alam, M. et al. Identification and quantification of Cr, Cu, and As incidental nanomaterials derived from CCA-treated wood in wildland-urban interface fire ashes. J. Hazard. Mater. 445, 130608 (2023).  
+
+<|ref|>text<|/ref|><|det|>[[174, 333, 870, 369]]<|/det|>
+"Following wildfires, severely burned areas are often barren and blanketed with ash and loose, rough topsoil leading to enhanced post-fire wind and water erosion \(^{8,39 - 42}\) ."  
+
+<|ref|>text<|/ref|><|det|>[[174, 385, 816, 421]]<|/det|>
+41. Yu, Y. & Ginoux, P. Enhanced dust emission following large wildfires due to vegetation disturbance. Nat. Geosci. 2022 1511 15, 878-884 (2022).  
+
+<|ref|>text<|/ref|><|det|>[[174, 437, 844, 473]]<|/det|>
+42. Shakesby, R. A. & Doerr, S. H. Wildfire as a hydrological and geomorphological agent. Earth-Science Rev. 74, 269-307 (2006).  
+
+<|ref|>text<|/ref|><|det|>[[174, 508, 876, 543]]<|/det|>
+"For example, CaCrO4 was noted after agricultural soil amended with composted Cr(III)- rich tannery sludge was heated at \(500^{\circ}\mathrm{C}^{55}\) ."  
+
+<|ref|>text<|/ref|><|det|>[[174, 560, 828, 612]]<|/det|>
+55. Rascio, I. et al. Evidence of hexavalent chromium formation and changes of Cr speciation after laboratory-simulated fires of composted tannery sludges long-term amended agricultural soils. J. Hazard. Mater. 436, 129117 (2022).
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[115, 91, 210, 108]]<|/det|>
+## Reviewer 2  
+
+<|ref|>text<|/ref|><|det|>[[115, 124, 876, 265]]<|/det|>
+The authors have to highlight the novelty of their manuscript. The abstract should be revised to attract the reader's attention. The Introduction section should be improved by adding references dealing soil contamination issues. However, the problem is that the English and the whole organisation of the present version are definitely below an acceptable standard for an international scientific journal. Analytical quality control is missing. Detection limits of the applied methods should be reported. The main problem for this manuscript is its structure. Major parts are missing from the manuscript. My suggestion is to reject this manuscript and encourage the authors to submit a more mature manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[173, 281, 870, 473]]<|/det|>
+Response: We appreciate the reviewer's feedback and have sought to make the abstract, and the manuscript, have more pizzazz. With that said, the manuscript was formatted specifically to the Nature guidelines. Further, the authors are all native speakers, and the senior author has published several hundred articles, including ones in Nature and Science. The present manuscript holds to those same standards. It is also worth noting that counter to Reviewer 2, Reviewer 1 stated "I found this paper very pleasant to read. The subject is well introduced, the sites and samples are well described and fit with the objective of the study, the results are well presented and they support the discussion...". Further, several established authors at Stanford have read the manuscript and all support the writing and presentation. Thus, while we don't want to dismiss the comments of the Reviewer, we do see them as anonymous in comparison to others.  
+
+<|ref|>text<|/ref|><|det|>[[173, 490, 877, 752]]<|/det|>
+Based on the reviewer's feedback, we have revised our Methods section to include more information regarding our analyses. We have added detection limits for \(\mathrm{Cr(VI)}\) and \(\mathrm{NH_4}\) measurements, and have revised related Figures and Tables to reflect these changes. With the exception of Na (detection limit of \(0.1\mathrm{mg / g}\) ), total concentrations for all elements reported using XRF were significantly greater than respective detection limits. We periodically analyzed certified reference material, NIST 2711a, with bulk soil samples to confirm accuracy of the XRF instrument. For aqueous extractions and associated instrument analyses, we tracked quality assurance in multiple ways. For each round of aqueous extractions using \(10\mathrm{mM}\mathrm{K_2HPO_4 / KH_2PO_4}\) solution, we included at least two centrifuge tubes containing the phosphate buffer solution and no soil/ash that were analyzed similar to samples for \(\mathrm{Cr(VI)}\) and \(\mathrm{NH_4}\) . Unless soil mass was limited, extractions were conducted in triplicate to assess sample heterogeneity. On the UV- Vis and ICP- MS, we analyzed instrument blanks every 15- 20 samples and multiple quality control standard solutions prepared with certified Cr reference solutions throughout each analysis.  
+
+<|ref|>text<|/ref|><|det|>[[174, 768, 875, 856]]<|/det|>
+Changes: We have modified the Abstract, added references to the Introduction section, and have sought to ensure the format and writing are consistent with the expected quality of the Nature journals. We have also revised the Methods to include detection limits, where relevant, and have similarly revised Figures and Tables to reflect non- detectable sample concentrations throughout the manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[175, 873, 510, 890]]<|/det|>
+Within the Methods section, we now state:
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[173, 90, 881, 457]]<|/det|>
+"Aqueous Extractions and Chemical AnalysisReactive \(\mathrm{Cr(VI)}\) concentrations (most available fraction, including dissolved and adsorbed \(\mathrm{Cr(VI)}\) ) in bulk soil ash samples (bulk and particle size fraction \(< 53 \mu \mathrm{m}\) ) and within soil cores were extracted with \(10 \mathrm{mM} \mathrm{K}_2\mathrm{HPO}_4 / \mathrm{KH}_2\mathrm{PO}_4\) (buffered at \(\mathrm{pH} 7.2\) ) \(^{79,80}\) . Phosphate effectively competes with \(\mathrm{Cr(VI)}\) ions for surface adsorption sites. At circumnatural to alkaline pH ranges in natural soils, it's expected that nearly all aqueous \(\mathrm{Cr}\) is present in the hexavalent form, and that \(\mathrm{Cr(VI)}\) concentrations will be primarily limited by adsorption \(^{81}\) . The clay size fraction (less than \(2 - \mu \mathrm{m}\) diameter) typically has a dominant influence on species retention given their high surface areas and greater number of adsorption sites; therefore, it is likely that the reactive \(\mathrm{Cr(VI)}\) concentrations measured here largely represent the fraction of \(\mathrm{Cr(VI)}\) associated with clay particles. Triplicate samples were agitated in a 1:4 soil/solution ratio for \(24 \mathrm{h}\) , centrifuged (30 min, \(4000 \mathrm{rpm}\) , \(4^{\circ}\mathrm{C}\) ), and filtered through \(0.22 - \mu \mathrm{m}\) filters. A subsample of unacidified filtrate was used to quantify aqueous \(\mathrm{Cr(VI)}\) concentrations using the diphenylcarbazide (DPC) method on a UV- Vis spectrophotometer (Shimadzu UV- 1601) \(^{79,82}\) . The detection limit was \(3 \mu \mathrm{g / L}\) (approximately \(12 \mu \mathrm{g / kg}\) ) \(^{82}\) . Total \(\mathrm{Cr}\) concentrations were determined with inductively coupled plasma mass spectrometry (ICP- MS, Thermo Scientific XSERIES 2), and confirmed that approximately all aqueous \(\mathrm{Cr}\) was in the form of \(\mathrm{Cr(VI)}\) in unburned soil and burned soil and ash, similarly observed in previous studies \(^{36,37}\) . An aliquot of each soil extract was immediately acidified post- filtration and stored in \(2\%\) nitric acid at \(4^{\circ}\mathrm{C}\) until ICP- MS analysis.  
+
+<|ref|>text<|/ref|><|det|>[[174, 472, 880, 613]]<|/det|>
+To determine relative differences in \(\mathrm{K}^+\) - extractable \(\mathrm{NH_4^+}\) concentrations (mg \(\mathrm{NH_4^+ - N / kg)}\) within burned and unburned soils (Figure S10), additional unacidified samples (after \(\mathrm{K_2HPO_4 / KH_2PO_4}\) extraction) from 30 of the 38 total soil cores (21 fire- affected and 9 unburned soil cores) were frozen at \(- 20^{\circ}\mathrm{C}\) until chemical analysis. Ammonium is a direct combustion product and will be elevated in the near surface soil after wildfires depending on burn severity \(^{40}\) . Ammonium concentrations in the top 6- cm were measured in triplicate (when sample volume allowed) using a flow injection analyzer (Westco SmartChem 200 Discrete Analyzer), with a detection limit of \(0.05 \mathrm{mg / L}\) \(^{82}\) .  
+
+<|ref|>sub_title<|/ref|><|det|>[[175, 630, 328, 647]]<|/det|>
+## Statistical Analyses  
+
+<|ref|>text<|/ref|><|det|>[[174, 664, 877, 805]]<|/det|>
+Means and standard errors were calculated for aqueous and solid- phase chemical measurements in all cores using replicates described below. Half the detection limit was used when measured concentrations were below detection limits. Total elemental concentrations were measured in 3- 4 solid- phase aliquots from each soil core (Table S1). At each soil depth interval (1- cm from 0- 6 cm; 2- cm from 6- 16 cm), triplicate aqueous extractions were conducted to evaluate reactive \(\mathrm{Cr(VI)}\) and exchangeable \(\mathrm{NH_4^+}\) concentrations (Figure 2, Figure 3a, Figure S10). In select soil depths within cores, replicates were limited (less than 3) due to solid mass or post- extraction aqueous volume.  
+
+<|ref|>text<|/ref|><|det|>[[174, 821, 875, 891]]<|/det|>
+To assess data normality, we applied the Shapiro- Wilk test and reported \(W\) statistics and p- values (Table S7). If data met normality assumptions at the \(95\%\) confidence interval (p- value \(= 0.05\) ), we used two- sided parametric tests; otherwise, we utilized two- sided nonparametric tests. Likewise, we used the f- test to determine equal variance. Unpaired \(t\)
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[173, 90, 867, 247]]<|/det|>
+tests were used to compare mean reactive Cr(VI) concentrations at the \(95\%\) confidence interval in near surface soil (0- 2 cm) of fire- affected and unburned sites based on geology. If one or both datasets were not normally distributed, such as in burned and unburned soils at control depths (10- 16 cm), Mann- Whitney U test was used. Within a soil core, we compared mean reactive Cr(VI) concentrations in surface soil (0- 2 cm) versus control depths (10- 16 cm) using either paired \(t\) test or Wilcoxon signed rank test. Detailed information about and results for each statistical analysis is provided in Tables S2- S5 and Table S7 the Supplementary Information. All statistical analyses were executed using the stats package in R (v. 4.1.3)."  
+
+<|ref|>sub_title<|/ref|><|det|>[[115, 281, 483, 300]]<|/det|>
+## From comments made on the manuscript pdf  
+
+<|ref|>text<|/ref|><|det|>[[115, 315, 866, 368]]<|/det|>
+Introduction section needs a short paragraph at the beginning to discuss elements distribution issues in wildfire impacted areas worldwide. More papers related to this paragraph will be beneficial for the paper.  
+
+<|ref|>text<|/ref|><|det|>[[115, 385, 566, 403]]<|/det|>
+May I suggest, among others, the following articles, e.g.:  
+
+<|ref|>text<|/ref|><|det|>[[115, 403, 850, 427]]<|/det|>
+1) Wildfire effects on soil quality. Application on a suburban area of West Attica (Greece).  
+
+<|ref|>text<|/ref|><|det|>[[115, 427, 780, 444]]<|/det|>
+Geosciences Journal, 25 (2), 243- 253 (https://doi.org/10.1007/s12303- 020- 0011- 1).  
+
+<|ref|>text<|/ref|><|det|>[[115, 444, 852, 477]]<|/det|>
+2) Suburban areas in flames: Dispersion of potentially toxic elements from burned vegetation and buildings. Estimation of the associated ecological and human health risk. Environmental Research, 183, 109153, https://doi.org/10.1016/j.envres.2020.109153.  
+
+<|ref|>text<|/ref|><|det|>[[115, 478, 850, 493]]<|/det|>
+Research, 183, 109153, https://doi.org/10.1016/j.envres.2020.109153.  
+
+<|ref|>text<|/ref|><|det|>[[115, 494, 850, 528]]<|/det|>
+3) Elements' Content in Stream Sediment and Wildfire Ash of Suburban Areas in West Attica (Greece). Water 2022, 14, 310. https://doi.org/10.3390/w14030310  
+
+<|ref|>text<|/ref|><|det|>[[173, 544, 855, 682]]<|/det|>
+Response: We thank the reviewer for the suggestion to add a paragraph discussing elemental concentrations of soil and ash after wildfires. We address past work that has quantified elemental concentrations in soils and ash in the first two paragraphs of the Introduction. We have added the second reference the reviewer suggested to our main text, in addition to a few recent studies on metals and their prevalence as a function of structural burning. References 1 and 3 address elemental concentrations within stream sediments of WUIs, and are ancillary to the introduction on metals within airborne particulate matter, and soil and ash, which are the focus of our study.  
+
+<|ref|>text<|/ref|><|det|>[[173, 699, 839, 733]]<|/det|>
+Changes: Based on both reviewers' recommendations, we have added the following references within the main text:  
+
+<|ref|>text<|/ref|><|det|>[[173, 750, 869, 785]]<|/det|>
+10. Xu, R. et al. Wildfires, Global Climate Change, and Human Health. N. Engl. J. Med. 383, 2173-2181 (2020).  
+
+<|ref|>text<|/ref|><|det|>[[173, 803, 860, 855]]<|/det|>
+26. Jahn, L. G. et al. Metallic and crustal elements in biomass-burning aerosol and ash: Prevalence, significance, and similarity to soil particles. ACS Earth Sp. Chem. 5, 136-148 (2021).
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[173, 90, 830, 126]]<|/det|>
+27. Ré, A. et al. Cytotoxic effects of wildfire ashes: In-vitro responses of skin cells. Environ. Pollut. 285, 117279 (2021).  
+
+<|ref|>text<|/ref|><|det|>[[173, 142, 861, 195]]<|/det|>
+29. Boaggio, K. et al. Beyond Particulate Matter Mass: Heightened Levels of Lead and Other Pollutants Associated with Destructive Fire Events in California. Environ. Sci. Technol. 56, 14272-14283 (2022).  
+
+<|ref|>text<|/ref|><|det|>[[173, 212, 870, 247]]<|/det|>
+30. Alshehri, T. et al. Wildland-urban interface fire ashes as a major source of incidental nanomaterials. J. Hazard. Mater. 443, 130311 (2023).  
+
+<|ref|>text<|/ref|><|det|>[[173, 264, 878, 316]]<|/det|>
+32. Alexakis, D. E. Suburban areas in flames: Dispersion of potentially toxic elements from burned vegetation and buildings. Estimation of the associated ecological and human health risk. Environ. Res. 183, 109153 (2020).  
+
+<|ref|>text<|/ref|><|det|>[[173, 333, 866, 386]]<|/det|>
+33. Alam, M. et al. Identification and quantification of Cr, Cu, and As incidental nanomaterials derived from CCA-treated wood in wildland-urban interface fire ashes. J. Hazard. Mater. 445, 130608 (2023).  
+
+<|ref|>text<|/ref|><|det|>[[173, 402, 818, 438]]<|/det|>
+41. Yu, Y. & Ginoux, P. Enhanced dust emission following large wildfires due to vegetation disturbance. Nat. Geosci. 2022 1511 15, 878-884 (2022).  
+
+<|ref|>text<|/ref|><|det|>[[173, 454, 844, 490]]<|/det|>
+42. Shakesby, R. A. & Doerr, S. H. Wildfire as a hydrological and geomorphological agent. Earth-Science Rev. 74, 269-307 (2006).  
+
+<|ref|>text<|/ref|><|det|>[[173, 507, 828, 560]]<|/det|>
+55. Rascio, I. et al. Evidence of hexavalent chromium formation and changes of Cr speciation after laboratory-simulated fires of composted tannery sludges long-term amended agricultural soils. J. Hazard. Mater. 436, 129117 (2022).  
+
+<|ref|>text<|/ref|><|det|>[[173, 576, 856, 612]]<|/det|>
+80. James, B. R. & Bartlett, R. J. Behavior of Chromium in Soils: VII. Adsorption and Reduction of Hexavalent Forms1. J. Environ. Qual. 12, 177 (1983).  
+
+<|ref|>text<|/ref|><|det|>[[173, 628, 841, 664]]<|/det|>
+81. Rai, D., Eary, L. E. & Zachara, J. M. Environmental chemistry of chromium. Sci. Total Environ. 86, 15-23 (1989).  
+
+<|ref|>text<|/ref|><|det|>[[173, 680, 857, 733]]<|/det|>
+82. McClain, C. N., Fendorf, S., Johnson, S. T., Menendez, A. & Maher, K. Lithologic and redox controls on hexavalent chromium in vadose zone sediments of California's Central Valley. Geochim. Cosmochim. Acta 265, 478-494 (2019).  
+
+<|ref|>text<|/ref|><|det|>[[116, 785, 875, 806]]<|/det|>
+Figure 2: Change A, B, and C in Figure plot panels to a, b, and c, reflected in the figure caption  
+
+<|ref|>text<|/ref|><|det|>[[171, 821, 848, 840]]<|/det|>
+Response: We appreciate the reviewer's feedback and have made the salient changes.  
+
+<|ref|>text<|/ref|><|det|>[[173, 856, 867, 891]]<|/det|>
+Changes: We have revised the figure panels (see below) in addition to references to the figure throughout the main text.
+
+<--- Page Split --->
+<|ref|>image<|/ref|><|det|>[[120, 110, 872, 360]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[115, 398, 714, 417]]<|/det|>
+Please avoid using "we". Apply this type of revision all over the manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[173, 433, 875, 486]]<|/det|>
+Response: We appreciate the reviewer's concern for the use of "we" within the manuscript. According to the editor and style guidelines for Nature Communications, the use of "we" is allowed and in some places encouraged.  
+
+<|ref|>text<|/ref|><|det|>[[173, 503, 447, 521]]<|/det|>
+Changes: No changes were made.  
+
+<|ref|>text<|/ref|><|det|>[[115, 538, 312, 555]]<|/det|>
+Please avoid using "we".  
+
+<|ref|>text<|/ref|><|det|>[[173, 572, 529, 590]]<|/det|>
+Response: Please refer to previous response.  
+
+<|ref|>text<|/ref|><|det|>[[173, 608, 447, 625]]<|/det|>
+Changes: No changes were made.  
+
+<|ref|>text<|/ref|><|det|>[[115, 642, 875, 661]]<|/det|>
+Figure 3: Change A, B, and C in Figure plot panels to a, b, and c, reflected in the figure caption  
+
+<|ref|>text<|/ref|><|det|>[[172, 677, 784, 696]]<|/det|>
+Response: We appreciate the reviewer's feedback agree with the suggestions.  
+
+<|ref|>text<|/ref|><|det|>[[172, 712, 867, 747]]<|/det|>
+Changes: We have revised the figure panels (see below) in addition to references to the figure throughout the main text.
+
+<--- Page Split --->
+<|ref|>image<|/ref|><|det|>[[122, 117, 868, 325]]<|/det|>
+<|ref|>image_caption<|/ref|><|det|>[[115, 373, 820, 393]]<|/det|>
+<center>Figure 4: Change A and B in Figure plot panels to a and b reflected in the figure caption </center>  
+
+<|ref|>text<|/ref|><|det|>[[171, 408, 818, 428]]<|/det|>
+Response: We appreciate the reviewer's feedback and agree with the suggestions.  
+
+<|ref|>text<|/ref|><|det|>[[173, 444, 867, 514]]<|/det|>
+Changes: We have revised the figure panels (see below) in addition to references to the figure throughout the main text. Please note that this figure was moved to the Supplementary Information (revised Figure S8), based on the recommendation of Reviewer 1.  
+
+<|ref|>image<|/ref|><|det|>[[130, 535, 867, 835]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[115, 860, 662, 878]]<|/det|>
+Where are the sections "Results and Discussion" and "Conclusions"?
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[173, 89, 866, 265]]<|/det|>
+Response: We have followed the format for Nature, which often do not have specific sections denoted as "Results and Discussion" nor "Conclusions". Further, within the author guidelines for Nature Communication submissions: "Nature Communications is flexible with regard to the format of initial submissions. Within reason, style and length will not directly influence consideration of a manuscript. We also do not require a particular structure or format at first submission. If and when revisions are required, the editor will provide detailed formatting instructions at that time." Based on the editor's instructions for manuscript revisions, we have defined "Introduction" and "Results" sections in the manuscript according to the Nature Communications formatting instructions.  
+
+<|ref|>text<|/ref|><|det|>[[174, 281, 855, 334]]<|/det|>
+Changes: Based on the editor's instructions for manuscript revisions, we have defined "Introduction" and "Results" sections in the manuscript according to the Nature Communications formatting instructions.
+
+<--- Page Split --->

peer_reviews/1254719c0acba0418d8fd83238729cf53dbf1fdc706c2f360fa1b3b6652b884b/supplementary_0_Transparent Peer Review file/images_list.json

@@ -0,0 +1 @@
+[]

peer_reviews/1254719c0acba0418d8fd83238729cf53dbf1fdc706c2f360fa1b3b6652b884b/supplementary_0_Transparent Peer Review file/supplementary_0_Transparent Peer Review file.mmd

@@ -0,0 +1,231 @@
+
+# nature portfolio  
+
+Peer Review File  
+
+# Temperate forests can deliver future wood demand and climate-change mitigation dependent on afforestation and circularity  
+
+Corresponding Author: Ms Eilidh Forster  
+
+This file contains all reviewer reports in order by version, followed by all author rebuttals in order by version.  
+
+Version 0:  
+
+Reviewer comments:  
+
+Reviewer #1  
+
+(Remarks to the Author)  
+
+The manuscript "Can temperate forests deliver both future wood demand and climate- change mitigation?" by Forster et al. presents the results of the lifecycle assessments of the global warming potential of wood supply under the growing wood demand scenarios. The analysis is fairly clear and thorough, however I have several reservations outlined below:  
+
+1. There are no uncertainty estimates around the wood supply and demand as well as GWP. Incorporating the uncertainty analysis into the study would provide insight into where the future research efforts should be directed in order to reduce the uncertainty, as well as help evaluate the confidence in the reported estimates. For an example of uncertainty analysis using CBM-CFS please refer to J.M. Metsaranta, C.H. Shaw, W.A. Kurz, C. Boisvenue, and S. Morken. 2017. Uncertainty of inventory-based estimates of the carbon dynamics of Canada's managed forest (1990-2014). Canadian Journal of Forest Research. 47(8): 1082-1094. https://doi.org/10.1139/cjfr-2017-0088  
+
+2. Although authors did state in the methods section that risks of pests disease, wind, fire, warming and other climate change effects would not significantly alter the study findings because the risks are highly uncertain and would apply similarly across the study, I don't think it is the case. Fire, drought, changing productivity due to climate change would not change the wood demand, however they could profoundly affect the wood supply. Given the length of the projections in the study, warming would substantially increase heterotrophic respiration, and therefore would alter GWP projections. Drought events could substantially and repeatedly reduce forest yield due via inhibition of the photosynthetic rate and increasing mortality rate. Warming temperatures and changing precipitation regimes are also likely to affect forest net primary production. Increasing fire frequency and severity could profoundly affect the wood supply. I don't think these effects are negligible for GWP and wood supply estimates, and therefore should be considered in the study, especially given the 100-year projection time.  
+
+3. The study results are generalized for a "temperate country", however most of the data used in the study are for the UK, why not make the study focused on the UK? It would make it easier to assess the feasibility of the proposed scenarios and allow to avoid generalized statements (as a reader, I had a little trouble with those).  
+
+Reviewer #2  
+
+(Remarks to the Author)  
+
+Summary: In this manuscript, authors present a combined application of forest carbon models and life cycle assessment (LCA) to estimate global warming (GW) impacts (+/-) of harvested wood products' (HWPs) value chain in a temperate country (appears to be UK). For projected high and low wood demands, the GW impacts are assessed for different scenarios within the country (changing rotation length of forests, increasing rate of production, and expansion of forest area) and overseas imports from non-temperate forests (tropical country). This kind of study is an interesting attempt to integrate two areas of the forest products value chain (forestry and HWPs) for policy implications. The results indicate that increased wood use is not a climate-change solution unless afforestation, increasing forest productivity under sustainable forest management, and mitigating demand increases through enhanced circularity and cascading of wood use are also integrated
+
+<--- Page Split --->
+
+into the strategy. I am offering below some comments/suggestions to improve the manuscript.  
+
+Major comments  
+
+103- 104: The current production and consumption levels of this temperate country should have been characterized to visualize the gap between demand and supply for the reference year 2023, and the how different scenarios or intended decisions might close this gap and influence GW impacts.  
+
+107- 108: What was the reason for selection of Sitka spruce forest in afforestation? The authors might discuss whether the results would be different if the forests were Douglas- fir or Western hemlock.  
+
+Figure 2: Please change y- axis units to Tg CO2e.  
+
+234: More interpretations could be added in 'high wood demand projection' results.  
+
+264- 267: Not clearly explained how GWP impact of alternative increased due to increase in overseas wood supply. Prolonged use of non- wood product and fuel alternatives?  
+
+278- 279: Please add explanation to this sentence, it seems confusing that more imports from tropical afforestation is better. 295: Clarity is needed on what type of non- wood product is considered substituted by HWPs. Because substitution credits for two type of non- wood products can be different for same HWP.  
+
+340: BECCS is associated with 'energy substitution' like non- wood products avoided is associated with 'product substitution'. There aren't a lot of interpretations in the results and discussion that focused on energy substitution. 600: there is no 'ix' in the components of LCA system boundary. Including via ...?  
+
+724- 740: This whole paragraph does not seem to be fit for methodology. Its more suitable for introduction.  
+
+747- 748: What is the per capita timber consumption in UK, which translates to \(30\%\) increase in demand by 2050? 754: In this paragraph, it would be good if authors give a brief about the different HWP end- uses (primary and cascading  
+
+uses) considered and maybe a justification for selecting HWP uses.  
+
+766: More clarity on defining the overseas forest type and carbon storage. Also, was the transportation distance and mode of transport included in the analysis?  
+
+785: Why 'GWP (forest C) impact...'? Forest carbon can be stored, emitted, or removed but cannot be equated directly to GWP.  
+
+Overall, the methods section appears weak to me and needs a thorough revision to ensure that work can be reproduced. Minor comments  
+
+325 and 330: should it be trip or tip?  
+
+723: why question marks in middle of the sentence?  
+
+Double numbering in references  
+
+Some of the links in supplementary excel file are broken. Please check.  
+
+## Reviewer #3  
+
+(Remarks to the Author)  
+
+The aim of this case study is to quantify the GHG mitigation potential of different measures or forest management options (in particular afforestation) in terms of meeting an increasing demand for wood, assuming that existing models for the forest sector tend to "underestimate" forest carbon fluxes.  
+
+In fact, in addition to the development of forest carbon stocks, the delayed release of biogenic carbon through the use of wood as material as well as potential shifting effects of the GHG emission balances associated with the life cycle of these wood- containing product systems and their potential product alternatives do also have an impact on the overall GHG balance.  
+
+However, the implementation of the presented approach to estimate the total GHG impact of different management scenarios compared to the defined reference, in our view appears to be completely inadequate in this study. It does not comply with applicable international standards and existing state- of- the- art knowledge.  
+
+While the modeling of carbon storage development in the forest using the internationally recognized Carbon Budget Model for the Canadian Forestry Service appears adequate, the methodological inadequacies relate in particular to the life cycle assessment methodology, on the basis of whose standard- compliant, consistent and transparent implementation avoided emissions through "product substitution" can be estimated in the first place. In order to adequately consider process chain emissions and based on those also "avoided emissions from product substitution", it is crucial to meet the internationally standardized requirements for life cycle assessment (including ISO 14040/44 and ISO 21930). The mere summation of unrelated LCA process information from a background database (here: Ecolnvent) is inadequate - at least for the processes outside the forest along the processing and value chain. The data used for calculating "primary avoided emissions (FF/product substitution)" are also completely unsuitable.  
+
+Furthermore, the simplified calculation of assumed carbon storage effects through the use of wood as a material (harvested wood products) contradicts central core requirements in the calculation as set out in the methodological guidelines and requirements provided by IPCC (incl. e.g. the consideration of inherited emissions). In consequence, the HWP contribution trough biogenic carbon storage as well as potential "avoided emissions" appear to be massively overestimated.  
+
+While it seems undisputed, even without the present study, that "the expansion of the (industrial) bioeconomy should be linked to the availability of raw materials in order to avoid unrealistic supply expectations," statements to the effect that "considerable HWP- C storage and product substitution credits can be achieved simultaneously" are not at all tenable on the basis of this simple and, in our view, completely methodologically inadequate implementation.
+
+<--- Page Split --->
+
+The simple comparison of a changing supply of forest wood with a modeled demand for this raw material (including potential gaps) can also be carried out without balancing all GHG emissions relevant to the forestry and wood sector.  
+
+Version 1:  
+
+Reviewer comments:  
+
+Reviewer #1  
+
+(Remarks to the Author)  
+
+The authors provided fairly comprehensive responses to the reviewers' comments. Yet, after reading the revised paper there is a dissonance between the level of detail in the manuscript and practical applicability of the findings. In the response to reviewers authors noted that this manuscript implements a framework laid out in an earlier paper published in Nature Communications, however the implementation is rather abstract, an interested party (e.g. government of a temperate country) would have to re- do the analyses (i.e. implement the framework to their specific country) and may get substantially different results. This makes me question the value of this particular manuscript, given lack of connection to any specific country, for which the feasibility of the generated estimates could be evaluated.  
+
+In my opinion, the treatment of the uncertainties associated with the effects of climate change and disturbances on productivity and GWP was fairly simplistic, not illustrated in all figures and not illustrated for GWP estimates. Lastly, please correct the following: CBM- CFS stands for "Carbon Budget Model of the Canadian Forest Sector".  
+
+Reviewer #2  
+
+(Remarks to the Author)  
+
+Open Access This Peer Review File is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.  
+
+In cases where reviewers are anonymous, credit should be given to 'Anonymous Referee' and the source.  
+
+The images or other third party material in this Peer Review File are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.  
+
+To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/
+
+<--- Page Split --->
+
+# Author Responses to Reviewer Comments
+
+We are grateful to and thank the reviewers for taking time to consider our manuscript.
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td></td><td>Reviewer #1:</td><td></td></tr><tr><td></td><td>The manuscript "Can temperate forests deliver both<br>future wood demand and climate-change mitigation?" by Forster et al. presents the results of the lifecycle<br>assessments of the global warming potential of wood<br>supply under the growing wood demand scenarios. The analysis is fairly clear and thorough, however I have<br>several reservations outlined below:</td><td>We thank the reviewer for the positive comments and valuable specific suggestions to<br>further enhance the paper.</td></tr><tr><td>1</td><td>There are no uncertainty estimates around the wood<br>supply and demand as well as GWP. Incorporating the<br>uncertainty analysis into the study would provide<br>insight into where the future research efforts should be directed in order to reduce the uncertainty, as well as<br>help evaluate the confidence in the reported estimates. For an example of uncertainty analysis using CBM-CFS<br>please refer to J.M. Metsaranta, C.H. Shaw, W.A. Kurz,<br>C. Boisvenue, and S. Morken. 2017. Uncertainty of<br>inventory-based estimates of the carbon dynamics of<br>Canada's managed forest (1990-2014). Canadian<br>Journal of Forest Research. 47(8): 1082-1094.<br>https://doi.org/10.1139/cjfr-2017-0088</td><td>The demand profiles modelled are based on the outcome of an extensive literature<br>review of wood demand modelling studies. We recognise and account for the significant<br>uncertainty sourrounding future wood demand projections (indicated by the wide range<br>of demand projections published in the literature) by modelling two different demand<br>profiles in the present study. These two demand profiles capture the range in the<br>published literature. LCA of forest management scenarios under each of these demand<br>profile senarios indicates the degree of uncertainty and sensitivity of results to future<br>demand projection assumptions. In dealing with uncertainties for the supply side, we<br>have taken the simple and clear approach of testing multiple scenarios of variation in<br>future forest productivity (yield). We believe that this is an efficient way of integrating all of these causes of uncertainty listed by the reviewer 1 comment 2. We have also added<br>the following text to the 'Discussion section' lines 375-379: "There is an urgent need for<br>more integrated evidence that incorporates holistic assessment of prospective forestry<br>value chains alongside landscape dynamics (including forest management and<br>expansion), at both national and global scales, including improved estimates of the<br>potential impacts of climate change-linked threats to the future productivity of both<br>temperate forests and the other forest biomes providing wood production."<br>See continuation of this response in response to reviewer 1 comment 2 below.</td></tr></table>
+
+<--- Page Split --->
+
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td>2</td><td>Although authors did state in the methods section that risks of pests disease, wind, fire, warming and other climate change effects would not significantly alter the study findings because the risks are highly uncertain and would apply similarly across the study, I don't think it is the case. Fire, drought, changing productivity due to climate change would not change the wood demand, however they could profoundly affect the wood supply. Given the length of the projections in the study, warming would substantially increase heterotrophic respiration, and therefore would alter GWP projections. Drought events could substantially and repeatedly reduce forest yield due via inhibition of the photosynthetic rate and increasing mortality rate. Warming temperatures and changing precipitation regimes are also likely to affect forest net primary production. Increasing fire frequency and severity could profoundly affect the wood supply. I don't think these effects are negligible for GWP and wood supply estimates, and therefore should be considered in the study, especially given the 100-year projection time.</td><td>As described in response to the previous comment, we believe that the approach to test multiple scenarios of variation in future forest productivity (yield) is an efficient way of integrating many of the causes of uncertainty listed by the reviewer. It is the approach that is most compatible with the LCA methodology of our study. Therefore, to address this specific reviewer requirement we have introduced a further pair of scenarios ('lower productivity' than the reference scenario) to model impact of possible ecosystem shocks (leading to yield reduction) that could arise from natural disturbance such as pests, disease, wind, drought and fire. These new scenarios are a modification of the 'reference rotation' existing forest (YC18, 50-yr rotation) + afforestation combination. The new scenarios are a transition of a) 15% and b) 30% of the existing forest from YC18 to YC12 ramping up over a 15 year period (remaining as a 50-yr harvest rotation), followed by recovery back to YC18 after one harvest rotation, ramping back down over a 15-yr period. (The afforestation assumptions were unchanged). Essentially the scenario represents a rapid shock of -two different intensities (affect on 15% &amp;amp; 30% of the total area), followed by recovery after harvest (at 50-yrs old). A detailed description of the basis for the parameterisation of these two scenarios (two and a half pages of text) based on a substantial new literature review (32 references cited) is provided in the new Supplementary Information file "Modelling of natural disturbances - 'reduced productivity'". We recognise that natural disturbances, the intensity and rate of impact, and management responses to these are complex and therefore the modelling options available are vast. We believe the new scenarios selected correspond to contrasting realistic disturbance patterns presented in the reviewed literature and therefore offer additional insights to the other scenarios already modelled. We have included the new results in an updated version of 'Supplementary Data 3', alongside the results of all the other scenarios modelled in this paper.<br>We modelled these two new scenarios under high and low demand projections and for the variation in afforestation rates and planting periods. We found that the new 'lower productivity' scenarios in this set led to higher net forest carbon loss (domestic and overseas) than the 'reference rotation' scenarios, as well as other changes, yet these disturbances do not change the overall results or findings of the study. We have added text on this to the methodology (lines 701-708) and results section of the manuscript</td></tr></table>
+
+<--- Page Split --->
+
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses<br>(lines 114-118; 169-172; and 201-204) and have included a more comprehensive description in the new Supplementary Information file (Supplementary Methods 1) in order to minimise additional word count.</td></tr></table>
+
+<--- Page Split --->
+
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td>3</td><td>The study results are generalized for a “temperate country”, however most of the data used in the study are for the UK, why not make the study focused on the UK? It would make it easier to assess the feasibility of the proposed scenarios and allow to avoid generalized statements (as a reader, I had a little trouble with those).</td><td>We intentionally kept the analysis generic for a hypothetical temperate country rather than a UK-specific case study in order to provide evidence that has broad relevance. (The UK is not 'typical' in that it has a high baseline of imports and low domestic wood production.) This study is prospective in nature and focuses on the impacts of change, based on transparent assumptions for future changes (e.g. rate of decarbonisation, projected wood demand increase, shifts in forest management, different afforestation rates) that are not unique to the UK (or any other country) or dependant on historic conditions that may be unique to the UK.<br>In order to perform the LCA we needed to make assumptions on typical product breakouts at the forest gate and at a sawmill and a representative wood product flow. Given limitations on public availability of this kind of data, we used data available for the UK as the basis for these assumptions. These assumptions are broadly representative across temperate countries beyond the UK given the similarities in technology and wood value chains. For example, typical rates of conversion of softwood to sawtimber at sawmills is similar across temperate regions. We have edited the Methodology 'Scope of LCA' section (e.g. lines 622-628) to try and improve the reconciliation between our use of UK data and generalised statements.</td></tr></table>
+
+<--- Page Split --->
+
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td></td><td>Reviewer #2</td><td></td></tr><tr><td></td><td>Summary: In this manuscript, authors present a combined application of forest carbon models and life cycle assessment (LCA) to estimate global warming (GW) impacts (+/-) of harvested wood products' (HWPs) value chain in a temperate country (appears to be UK). For projected high and low wood demands, the GW impacts are assessed for different<br>scenarios within the country (changing rotation length of forests, increasing rate of production, and expansion of forest area) and overseas imports from non-temperate forests (tropical country). This kind of study is an interesting attempt to integrate two areas of the forest products value chain (forestry and HWPs) for policy implications. The results indicate that increased wood use is not a climate-change solution unless afforestation, increasing forest productivity under sustainable forest management, and mitigating<br>demand increases through enhanced circularity and<br>cascading of wood use are also integrated into the strategy. I am offering below some comments/suggestions to improve the manuscript.</td><td>We thank the reviewer for the positive comments and endorsement of the important findings of the paper.</td></tr><tr><td></td><td>Major comments</td><td></td></tr><tr><td>1</td><td>103-104: The current production and consumption levels of this temperate country should have been characterized to visualize the gap between demand and supply for the<br>reference year 2023, and the how different scenarios or<br>intended decisions might close this gap and influence GW<br>impacts.</td><td>We intentionally defined a hypothetical scenario in which domestic production equals domestic consumption in the reference year 2023, as indicated in Figure 1. As described in our response to reviewer 1 comment 3 above, we have taken this hypothetical approach rather than characterising the actual production and<br>consumption levels of a specific country in order to provide evidence that has<br>broad relevance. (For example, the UK is not 'typical' in that it has a high baseline of imports and low domestic wood production.) It makes sense for a<br>counterfactual lifecycle assessment (which is concerned with change from a<br>baseline) to start the scenario with production = consumption rather that starting</td></tr></table>
+
+<--- Page Split --->
+
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td></td><td></td><td>with a deficit or surplus, which would make interpretation of results more challenging.<br>We then visualise how the different scenarios close the gap between future (increasing) demand and supply in Figure 1 by colour coding the supply deficit in red (and the supply surplus in pink). We have also added some additional text in the Figure 1 title to improve this description.</td></tr><tr><td>2</td><td>107-108: What was the reason for selection of Sitka spruce forest in afforestation? The authors might discuss whether the results would be different if the forests were Douglas-fir or Western hemlock.</td><td>Sitka spruce was selected as it grows widely (native to west coast of Canada and the United States; now planted in 16 countries worldwide, including as the predominant plantation species in the UK, Ireland and Denmark - all countries with low forest cover). Sitka spruce also grows well in degraded upland sites where land is most economically viable for afforestation, unlike productive Douglas fir or western hemlock, which require lower altitude land that is of higher value for food production. LCA results would not be different if different species (or species mixes) were modelled unless a significantly different yield class (growth curve) was also assumed. It is the yield rather than the species choice that drives the GWP impact results (which of course would be related in specific geographic contexts but in a hypothetical forest study such as this, the yield class is specified independently by the modeller as a representative average across contexts). In this way, the choice of Sitka spruce is simply a proxy for commercial conifer species. However, we believe that the clarity of the manuscript benefits from naming the case species, which links to the UK sawmill and-representative wood product flow data used in the study.</td></tr><tr><td>3</td><td>Figure 2: Please change y-axis units to Tg CO2e.</td><td>We have made this change.</td></tr><tr><td>4</td><td>234: More interpretations could be added in 'high wood demand projection' results.</td><td>We are pleased that Reviewer 2 sees further possible interpretations that we could include. We would welcome the opportunity to add more interpretations, however because of the journal's strict word limits this would necessitate making significant cuts to other parts of the text, which would sacrifice their clarity and content. Given the importance of these other components of the paper that would have to be cut, with regret we have concluded that adding more</td></tr></table>
+
+<--- Page Split --->
+
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td></td><td></td><td>interpretations to the 'high wood demand projection' results would be of net cost to the value of the paper.</td></tr><tr><td>5</td><td>264-267: Not clearly explained how GWP impact of alternative increased due to increase in overseas wood supply. Prolonged use of non-wood product and fuel alternatives?</td><td>We have added words in this sentence (line 290) to clarify that the GWP impact of the alternative is "curtailed HWP supply" i.e. prolonged use of non-wood product and fuel alternatives. We have also added further description to indicate the relative impact of the specific overseas scenarios ('Boreal 1,2&3, Tropical CVL and Tropical (afforestation)' in Fig. 4) in which emissions associated with supplying the wood shortfall are higher or lower that the emissions from 'curtailed HWP supply' (prolonged use of non-wood and fuel alternatives).</td></tr><tr><td>6</td><td>278-279: Please add explanation to this sentence, it seems confusing that more imports from tropical afforestation is better.</td><td>It is unclear to us why this is confusing as it seems to be clearly shown in the results reported in Fig. 4. We do acknowledge the potential disbenefits of new tropical afforestation in lines 363-364 ("numerous socio-economic51 and biodiversity conservation20 caveats". In essence, increasing demand for fast-growing tropical tree species, which could be established on the large areas of degraded land in the tropics (if local socio-economic conditions make them available), could actually enhance terrestrial carbon stocks.</td></tr><tr><td>7</td><td>295: Clarity is needed on what type of non-wood product is considered substituted by HWPs. Because substitution credits for two type of non-wood products can be different for same HWP.</td><td>In response to this point, to avoid adding significant additional word count, we have added text to the methodology (line 616-617) and in the Fig. 5 title to direct readers to supplementary information for more detail on the substitution credits assumptions (supplementary figure 1 and supplementary Table 1). We also direct readers to Forster et al. (2021) where the methodology was originally published. We have also improved figure 5 in the present paper to convey, among other things, the product substitution assumptions more clearly.</td></tr><tr><td>8</td><td>340: BECCS is associated with 'energy substitution' like non-wood products avoided is associated with 'product substitution'. There aren't a lot of interpretations in the results and discussion that focused on energy substitution.</td><td>We have expanded the sentence referring to BECCS to mention product substitution and emphasise the permanent geological storage of biogenic carbon that contributes a carbon sink. Lines 368-369.</td></tr></table>
+
+<--- Page Split --->
+
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td>9</td><td>600: there is no ‘ix’ in the components of LCA system boundary. Including via ...?</td><td>We have checked the numbering of components of the LCA system boundary in figure 5 and in the text and cannot find anything missing.</td></tr><tr><td>10</td><td>724-740: This whole paragraph does not seem to be fit for methodology. Its more suitable for introduction.</td><td>We believe that the paragraph is important for justifying the important selection of assumptions for the projected rate of demand increase used in the study, which is a core aspect of the study's methodology. We have edited and changed the order of the text to make the inclusion of this text in the methodology more fluent and justified. E.g. lines 779-785 have been moved from earlier in the section.</td></tr><tr><td>11</td><td>747-748: What is the per capita timber consumption in UK, which translates to 30% increase in demand by 2050?</td><td>We do not believe that including per capita timber consumption in the UK is relevant to the study. The 30% increase by 2050 statistic was included as this is a date that many published projections focus on, because it is an important date for many 'net zero' targets as mandated by the 2015 Paris Agreement of UNFCCC, and therefore a statistic that readers are likely to be interested in. As stated in our response to reviewer 2 comment 1, we have prioritised assumptions that are non-country-specific and forward looking wherever possible, to maximise the transferability of results. In essence, the critical (and transferable) mathematical/biophysical dynamic is the rate of demand increase from the baseline relative to the expansion rate of baseline forest area (through afforestation) and the (change in) productivity of existing (and new) forest.</td></tr><tr><td>12</td><td>754: In this paragraph, it would be good if authors give a brief about the different HWP end-uses (primary and cascading uses) considered and maybe a justification for selecting HWP uses.</td><td>The primary HWP end-uses are described in the methodology and further in the SI, including in a revised system diagram (Figure 5) indicating major processes and products. Details of cascading uses and also justification for the assumptions used are provided in Forster et al. (2021), which we cite in the methodology. We believe there is insufficient justification to repeat all of this detail here, especially on account of the word count restrictions.</td></tr><tr><td>13</td><td>766: More clarity on defining the overseas forest type and carbon storage. Also, was the transportation distance and mode of transport included in the analysis?</td><td>Description of the overseas forest types is provided in the following paragraphs (lines 827-856).<br>We did not include assumed differences in transport distances between overseas and domestic HWPs because, in a previous study (Forster et al., 2023), we found that these differences made a very small contribution (less than 0.01%) relative to the net GWP impact of the value chain.</td></tr></table>
+
+<--- Page Split --->
+
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td>14</td><td>785: Why 'GWP (forest C) impact...'? Forest carbon can be stored, emitted, or removed but cannot be equated directly to GWP.</td><td>GWP (forest C) impact is defined in the previous paragraph as the 'GWP impact of fluxes in forest carbon stocks'. We have added '(GWP (forest C))' after this definition to add clarity. It refers to the GWP impact associated with net CO2e fluxes to/from the forest. Line 828.</td></tr><tr><td>15</td><td>Overall, the methods section appears weak to me and needs a thorough revision to ensure that work can be reproduced.</td><td>We have carefully addressed each of the reviewers specific comments and assume that in doing so we have met the need expressed here.</td></tr><tr><td></td><td>Minor comments</td><td></td></tr><tr><td>16</td><td>325 and 330: should it be trip or tip?</td><td>Tip means to tilt, tumble or topple. In this context, 'the balance... can tip from sink to source', means the balance can tilt or shift from sink to source.</td></tr><tr><td>17</td><td>723: why question marks in middle of the sentence?</td><td>We have removed these.</td></tr><tr><td>18</td><td>Double numbering in references</td><td>We have removed these.</td></tr><tr><td>19</td><td>Some of the links in supplementary excel file are broken. Please check.</td><td>We have checked the supplementary Excel files and assume that the reviewer is referring to the 'HWP calculation module' (Supplementary Data 2). The broken links are the emissions factors cells that link to the original LCA worksheet (Supplementary Data 1) that generated them. These links do not work once the workbook is resaved in a different location. However, in the final version of these supplementary data files, which will be stored in a permanent repository or in the publisher's website, we will endeavour to ensure the links all work.</td></tr><tr><td></td><td></td><td></td></tr></table>
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td></td><td>Reviewer #3:</td><td></td></tr><tr><td>1</td><td>The aim of this case study is to quantify the GHG mitigation potential of different measures or forest management<br>options (in particular afforestation) in terms of meeting an<br>increasing demand for wood, assuming that existing models for the forest sector tend to "underestimate" forest carbon fluxes.</td><td></td></tr></table>
+
+<--- Page Split --->
+
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td>2</td><td>In fact, in addition to the development of forest carbon stocks, the delayed release of biogenic carbon through the use of wood as material as well as potential shifting effects of the GHG emission balances associated with the life cycle of these wood-containing product systems and their potential product alternatives do also have an impact on the overall GHG balance.</td><td>We are not clear on whether this forms part of the reviewer's summary of the paper or is a comment requiring our response. In case it is the latter, we provide the following clarification. We agree with the reviewer about the importance of impacts throughout the wood products life cycle. Yet, the full climate mitigation effects of forestry are often under-represented because: (i) LCA approaches that (sometimes) better represent substitution and possible end-of-life carbon storage effects are usually only applied to specific products (partial wood flows out of forests); (ii) inventory approaches that attempt to represent all wood products, typically neglect short-lived wood products used for energy, and don't attribute substitution effects back to the forest sector; (iii) both LCA and inventory approaches often disregard cascading uses of wood, and associated second (and possible further) substitutions, along with extended carbon storage effects; (iv) only very prospective analyses consider future BECCS deployment that could lock up biogenic carbon indefinitely.</td></tr><tr><td>3</td><td>However, the implementation of the presented approach to estimate the total GHG impact of different management scenarios compared to the defined reference, in our view appears to be completely inadequate in this study. It does not comply with applicable international standards and existing state-of-the-art knowledge.</td><td>We are surprised by this comment, which we strongly believe does not reflect the content and rigour of our study or its compliance with the highest international standards. International standards apply to specific types of accounting, such as Environmental Product Declarations for (LCA of) wood products or UNFCCC guidelines for national GHG inventory accounting. In contrast, there are no such "standards" for the prospective consequential LCA of entire forest and multiple downstream product systems, which is the innovative methodological approach of our study. That is precisely what makes this approach so valuable - it transcends the accounting rules that are applied to deal with truncated system boundaries (e.g. how to allocate forest effects to specific downstream products). This paper builds on the state-of-the-art approach applied in our earlier, highly-cited manuscript published in Nature Communications (Forster et al., 2021), and applies it to a temperate forest system under different management and expansion regimes; this is highly novel. To reiterate, we are not attempting to calculate product footprints as per various international standards, but instead (as is befitting for a submission to Nature Communications) to provide a novel, holistic and rigorous analysis of the climate mitigation efficacy of different forest</td></tr></table>
+
+<--- Page Split --->
+
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td></td><td></td><td>management strategies, providing key new evidence for policy makers and an innovative novel approach for researchers.</td></tr><tr><td>4</td><td>While the modeling of carbon storage development in the forest using the internationally recognized Carbon Budget Model for the Canadian Forestry Service appears adequate, the methodological inadequacies relate in particular to the life cycle assessment methodology, on the basis of whose standard-compliant, consistent and transparent implementation avoided emissions through "product substitution" can be estimated in the first place. In order to adequately consider process chain emissions and based on those also "avoided emissions from product substitution", it is crucial to meet the internationally standardized requirements for life cycle assessment (including ISO 14040/44 and ISO 21930). The mere summation of unrelated LCA process information from a background database (here: Ecolnvent) is inadequate - at least for the processes outside the forest along the processing and value chain. The data used for calculating "primary avoided emissions (FF/product substitution)" are also completely unsuitable.</td><td>As explained in our previous response, our study is categorically not an attributional LCA, which seems to have been the assumption of the reviewer. Instead it is a consequential LCA. That said, this study does fully comply with ISO 14040 and 14044 standards (which in reality are a basic framework to structure LCA), insofar as it: provides a clear goal and scope of the study; calculates an inventory that accounts for relevant changes within the expanded system boundary; applies GWP100 characterisation factors for the life cycle impact assessment phase; interprets results with clear regard for the question and methods used (now including important sensitivity analyses). Furthermore, this study uses system expansion in preference to allocation, as recommended in the allocation hierarchy advocated by the ISO standards. The study categorically does not sum "unrelated" LCA process information - incurred or avoided processes are clearly indicated in the system boundary (Supplementary Fig. 1 and in Fig. 5, also now improved), based on a consequential LCA approach that incorporates all changes associated with the scenarios vs the baseline counterfactual situation.</td></tr></table>
+
+<--- Page Split --->
+
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td>5</td><td>Furthermore, the simplified calculation of assumed carbon storage effects through the use of wood as a material (harvested wood products) contradicts central core requirements in the calculation as set out in the methodological guidelines and requirements provided by IPCC (incl. e.g. the consideration of inherited emissions). In consequence, the HWP contribution trough biogenic carbon storage as well as potential “avoided emissions” appear to be massively overestimated.</td><td>Our methodology does not contradict core requirements of the IPCC and from this comment we do not understand how the reviewer considers that it does so. The IPCC provide guidelines for a number of approaches to deal with HWP C storage at a national scale. All of them are based on the fact that biogenic carbon harvested from the forest is "lost" from the terrestrial system, but some of this carbon "reappears" in longer-lived HWPs (with a large deficit reflecting biomass assumed to be immediately oxidised via combustion for bioenergy or for kiln drying of wood in sawmills). Our approach fully respects the biogenic carbon balance - in fact much more explicitly than typical national inventory accounting (with which we are fully familiar), because flows into all main products, both short- and long-lived, are accounted for, and the duration of carbon storage in each of these products, and any subsequent products, is explicitly accounted for in our methodology. Again, this is the benefit of our expanded boundary approach (across products and through time) - it minimises the influence of "cut-off" rules and associated value judgements, which often profoundly influence carbon footprints at the wood product level.</td></tr><tr><td>6</td><td>While it seems undisputed, even without the present study, that "the expansion of the (industrial) bioeconomy should be linked to the availability of raw materials in order to avoid unrealistic supply expectations, " statements to the effect that "considerable HWP-C storage and product substitution credits can be achieved simultaneously" are not at all tenable on the basis of this simple and, in our view, completely methodologically inadequate implementation.</td><td>Without any further insight from the reviewer as to where our methodology is inadequate, it is difficult to address this comment. We acknowledge that our original system diagram did not fully represent the scope of our methodology (although it was described fully in the text and illustrated in Supplementary Fig.1). Therefore, we hope that the improved system diagram in the resubmitted manuscript helps understanding of precisely how we have linked forest C storage with HWP-C storage and substitution, and long-term CCS C-storage effects in a coherent manner. As previously mentioned, this builds on the methodology of Forster et al. (2021) published in Nature Communications.</td></tr><tr><td>7</td><td>The simple comparison of a changing supply of forest wood with a modeled demand for this raw material (including potential gaps) can also be carried out without balancing all GHG emissions relevant to the forestry and wood sector.</td><td>It is true that a simple comparison of changing supply of forest wood with demand could be carried out in isolation. However, that would not address the important policy question of what the climate mitigation effect would be of alternative forest strategies that aim to minimise the gap between future supply and demand. That is where this paper provides a unique and robust contribution to the scientific literature, for which its rigorous methodological consequential LCA approach is essential.</td></tr></table>
+
+<--- Page Split --->
+
+# 1.1.1.1.1.1.1.1.1.1.1
+
+<--- Page Split --->
+
+## Author response to reviewer comments  
+
+We are once again grateful for the opportunity to respond to Reviewer comments. We have addressed the comments below and in doing so believe we have further enhanced the article, as outlined in our response. All changes to the manuscript are highlighted in 'tracked- changes'.  
+
+## REVIEWER COMMENTS  
+
+Reviewer #1 (Remarks to the Author):  
+
+## Comment 1  
+
+The authors provided fairly comprehensive responses to the reviewers' comments. Yet, after reading the revised paper there is a dissonance between the level of detail in the manuscript and practical applicability of the findings. In the response to reviewers authors noted that this manuscript implements a framework laid out in an earlier paper published in Nature Communications, however the implementation is rather abstract, an interested party (e.g. government of a temperate country) would have to re- do the analyses (i.e. implement the framework to their specific country) and may get substantially different results. This makes me question the value of this particular manuscript, given lack of connection to any specific country, for which the feasibility of the generated estimates could be evaluated.  
+
+## Response to Comment 1  
+
+We thank Reviewer 1 for their further comments, to which we have given careful consideration. The main point raised here is that the methodology remains abstract, such that others would need to perform their own analysis. We fully agree that researchers and other interested parties should perform context- specific analysis, and not rely solely on the results of a generic country case study (of the kind we use here to demonstrate the rigour, power and applicability of the framework that we have developed in this paper). We have subtly revised framing to reflect this, with additional clarity and emphasis on the novel methodological framework we have developed. In a new SI, we outline the framework in diagram, table and text form, highlighting its substantial novel components not previously published, and describe its implementation with reference to important parameters and examples of datasets and models of the kind that are available to generate context- specific results. However, we maintain that using a generic temperate country with an even- aged forest as an illustrative case study provides important insights that do have broad relevance for readers, whereas using the context of a specific country would not, owing to the unique nature of its forest- age deviations. We have made small edits throughout the manuscript to further emphasise to readers the importance of tailoring studies to specific contexts, especially where the results of such studies are intended to inform policy.  
+
+The framework sets out our exploratory approach for identifying 'low regrets' climate solutions for the forestry value chain by evaluating a range of plausible future scenarios, using powerful integrated forest modelling and lifecycle assessment (LCA) (Supplementary Fig. 1 & Supplementary Methods 1).  
+
+The present study builds substantially on the detailed prospective dynamic LCA modelling of entire forest- wood value chains developed and explained in Forster et al. (2021). Crucially, in the present study the full downstream greenhouse gas (GHG) mitigation consequences of wood use and end- of
+
+<--- Page Split --->
+
+life management are considered - including carbon storage and material and energy substitution within a decarbonising future economy. Novel aspects to the present study (shown in pink in the new Supplementary Fig. 1) include calculating potential future wood supply deficit by comparing projected wood demand curves to wood supply from a range of modelled (expanded) temperate forest management scenarios, and linking the supply deficit to marginal expansion of supply from other regions.  
+
+Elaboration of the framework within the SI will facilitate interested parties to follow our approach for a relevant study context (Supplementary Table 1). We signpost readers to this SI at multiple points in the manuscript.  
+
+## Comment 2  
+
+In my opinion, the treatment of the uncertainties associated with the effects of climate change and disturbances on productivity and GWP was fairly simplistic, not illustrated in all figures and not illustrated for GWP estimates.  
+
+## Response to Comment 2  
+
+We acknowledge the simple treatment of climate change and disturbances (which we note can vary considerably, geographically) for our generic temperate country study. However, we maintain that this approach is appropriate to illustrate the magnitude of effect associated with highly uncertain future events. Recent articles pertaining to climate mitigation modelling highlight the need for "robust" rather than optimised decision making in the face of deep uncertainty and high complexity (Workman et al. (2021) and Workman et al. (2024)). We now cite these articles to further justify our approach, and propose natural disturbance events as a key parameter to represent (where possible) this impact within the proposed framework when tailoring scenarios in future studies for specific context.  
+
+## Comment 3  
+
+Lastly, please correct the following: CBM- CFS stands for "Carbon Budget Model of the Canadian Forest Sector".  
+
+## Response to Comment 3  
+
+Corrections made.
+
+<--- Page Split --->

peer_reviews/1254719c0acba0418d8fd83238729cf53dbf1fdc706c2f360fa1b3b6652b884b/supplementary_0_Transparent Peer Review file/supplementary_0_Transparent Peer Review file_det.mmd

@@ -0,0 +1,307 @@
+<|ref|>title<|/ref|><|det|>[[73, 53, 295, 80]]<|/det|>
+# nature portfolio  
+
+<|ref|>text<|/ref|><|det|>[[75, 97, 296, 119]]<|/det|>
+Peer Review File  
+
+<|ref|>title<|/ref|><|det|>[[73, 161, 874, 235]]<|/det|>
+# Temperate forests can deliver future wood demand and climate-change mitigation dependent on afforestation and circularity  
+
+<|ref|>text<|/ref|><|det|>[[73, 249, 420, 266]]<|/det|>
+Corresponding Author: Ms Eilidh Forster  
+
+<|ref|>text<|/ref|><|det|>[[70, 299, 864, 314]]<|/det|>
+This file contains all reviewer reports in order by version, followed by all author rebuttals in order by version.  
+
+<|ref|>text<|/ref|><|det|>[[73, 351, 144, 364]]<|/det|>
+Version 0:  
+
+<|ref|>text<|/ref|><|det|>[[73, 377, 219, 390]]<|/det|>
+Reviewer comments:  
+
+<|ref|>text<|/ref|><|det|>[[73, 403, 160, 416]]<|/det|>
+Reviewer #1  
+
+<|ref|>text<|/ref|><|det|>[[73, 430, 238, 443]]<|/det|>
+(Remarks to the Author)  
+
+<|ref|>text<|/ref|><|det|>[[73, 443, 904, 482]]<|/det|>
+The manuscript "Can temperate forests deliver both future wood demand and climate- change mitigation?" by Forster et al. presents the results of the lifecycle assessments of the global warming potential of wood supply under the growing wood demand scenarios. The analysis is fairly clear and thorough, however I have several reservations outlined below:  
+
+<|ref|>text<|/ref|><|det|>[[72, 494, 916, 574]]<|/det|>
+1. There are no uncertainty estimates around the wood supply and demand as well as GWP. Incorporating the uncertainty analysis into the study would provide insight into where the future research efforts should be directed in order to reduce the uncertainty, as well as help evaluate the confidence in the reported estimates. For an example of uncertainty analysis using CBM-CFS please refer to J.M. Metsaranta, C.H. Shaw, W.A. Kurz, C. Boisvenue, and S. Morken. 2017. Uncertainty of inventory-based estimates of the carbon dynamics of Canada's managed forest (1990-2014). Canadian Journal of Forest Research. 47(8): 1082-1094. https://doi.org/10.1139/cjfr-2017-0088  
+
+<|ref|>text<|/ref|><|det|>[[71, 586, 920, 716]]<|/det|>
+2. Although authors did state in the methods section that risks of pests disease, wind, fire, warming and other climate change effects would not significantly alter the study findings because the risks are highly uncertain and would apply similarly across the study, I don't think it is the case. Fire, drought, changing productivity due to climate change would not change the wood demand, however they could profoundly affect the wood supply. Given the length of the projections in the study, warming would substantially increase heterotrophic respiration, and therefore would alter GWP projections. Drought events could substantially and repeatedly reduce forest yield due via inhibition of the photosynthetic rate and increasing mortality rate. Warming temperatures and changing precipitation regimes are also likely to affect forest net primary production. Increasing fire frequency and severity could profoundly affect the wood supply. I don't think these effects are negligible for GWP and wood supply estimates, and therefore should be considered in the study, especially given the 100-year projection time.  
+
+<|ref|>text<|/ref|><|det|>[[72, 728, 909, 768]]<|/det|>
+3. The study results are generalized for a "temperate country", however most of the data used in the study are for the UK, why not make the study focused on the UK? It would make it easier to assess the feasibility of the proposed scenarios and allow to avoid generalized statements (as a reader, I had a little trouble with those).  
+
+<|ref|>text<|/ref|><|det|>[[73, 794, 161, 806]]<|/det|>
+Reviewer #2  
+
+<|ref|>text<|/ref|><|det|>[[73, 820, 238, 833]]<|/det|>
+(Remarks to the Author)  
+
+<|ref|>text<|/ref|><|det|>[[72, 833, 923, 938]]<|/det|>
+Summary: In this manuscript, authors present a combined application of forest carbon models and life cycle assessment (LCA) to estimate global warming (GW) impacts (+/-) of harvested wood products' (HWPs) value chain in a temperate country (appears to be UK). For projected high and low wood demands, the GW impacts are assessed for different scenarios within the country (changing rotation length of forests, increasing rate of production, and expansion of forest area) and overseas imports from non-temperate forests (tropical country). This kind of study is an interesting attempt to integrate two areas of the forest products value chain (forestry and HWPs) for policy implications. The results indicate that increased wood use is not a climate-change solution unless afforestation, increasing forest productivity under sustainable forest management, and mitigating demand increases through enhanced circularity and cascading of wood use are also integrated
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[72, 47, 707, 61]]<|/det|>
+into the strategy. I am offering below some comments/suggestions to improve the manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[72, 61, 188, 72]]<|/det|>
+Major comments  
+
+<|ref|>text<|/ref|><|det|>[[72, 73, 895, 112]]<|/det|>
+103- 104: The current production and consumption levels of this temperate country should have been characterized to visualize the gap between demand and supply for the reference year 2023, and the how different scenarios or intended decisions might close this gap and influence GW impacts.  
+
+<|ref|>text<|/ref|><|det|>[[72, 112, 900, 138]]<|/det|>
+107- 108: What was the reason for selection of Sitka spruce forest in afforestation? The authors might discuss whether the results would be different if the forests were Douglas- fir or Western hemlock.  
+
+<|ref|>text<|/ref|><|det|>[[73, 139, 409, 152]]<|/det|>
+Figure 2: Please change y- axis units to Tg CO2e.  
+
+<|ref|>text<|/ref|><|det|>[[72, 152, 641, 165]]<|/det|>
+234: More interpretations could be added in 'high wood demand projection' results.  
+
+<|ref|>text<|/ref|><|det|>[[72, 165, 860, 191]]<|/det|>
+264- 267: Not clearly explained how GWP impact of alternative increased due to increase in overseas wood supply. Prolonged use of non- wood product and fuel alternatives?  
+
+<|ref|>text<|/ref|><|det|>[[72, 191, 910, 229]]<|/det|>
+278- 279: Please add explanation to this sentence, it seems confusing that more imports from tropical afforestation is better. 295: Clarity is needed on what type of non- wood product is considered substituted by HWPs. Because substitution credits for two type of non- wood products can be different for same HWP.  
+
+<|ref|>text<|/ref|><|det|>[[72, 229, 846, 268]]<|/det|>
+340: BECCS is associated with 'energy substitution' like non- wood products avoided is associated with 'product substitution'. There aren't a lot of interpretations in the results and discussion that focused on energy substitution. 600: there is no 'ix' in the components of LCA system boundary. Including via ...?  
+
+<|ref|>text<|/ref|><|det|>[[72, 268, 792, 281]]<|/det|>
+724- 740: This whole paragraph does not seem to be fit for methodology. Its more suitable for introduction.  
+
+<|ref|>text<|/ref|><|det|>[[72, 281, 850, 308]]<|/det|>
+747- 748: What is the per capita timber consumption in UK, which translates to \(30\%\) increase in demand by 2050? 754: In this paragraph, it would be good if authors give a brief about the different HWP end- uses (primary and cascading  
+
+<|ref|>text<|/ref|><|det|>[[72, 308, 540, 320]]<|/det|>
+uses) considered and maybe a justification for selecting HWP uses.  
+
+<|ref|>text<|/ref|><|det|>[[72, 321, 920, 346]]<|/det|>
+766: More clarity on defining the overseas forest type and carbon storage. Also, was the transportation distance and mode of transport included in the analysis?  
+
+<|ref|>text<|/ref|><|det|>[[72, 347, 899, 373]]<|/det|>
+785: Why 'GWP (forest C) impact...'? Forest carbon can be stored, emitted, or removed but cannot be equated directly to GWP.  
+
+<|ref|>text<|/ref|><|det|>[[72, 373, 899, 399]]<|/det|>
+Overall, the methods section appears weak to me and needs a thorough revision to ensure that work can be reproduced. Minor comments  
+
+<|ref|>text<|/ref|><|det|>[[72, 399, 320, 412]]<|/det|>
+325 and 330: should it be trip or tip?  
+
+<|ref|>text<|/ref|><|det|>[[72, 413, 430, 425]]<|/det|>
+723: why question marks in middle of the sentence?  
+
+<|ref|>text<|/ref|><|det|>[[72, 426, 297, 437]]<|/det|>
+Double numbering in references  
+
+<|ref|>text<|/ref|><|det|>[[72, 438, 560, 451]]<|/det|>
+Some of the links in supplementary excel file are broken. Please check.  
+
+<|ref|>sub_title<|/ref|><|det|>[[72, 476, 162, 490]]<|/det|>
+## Reviewer #3  
+
+<|ref|>text<|/ref|><|det|>[[72, 503, 237, 515]]<|/det|>
+(Remarks to the Author)  
+
+<|ref|>text<|/ref|><|det|>[[72, 515, 916, 555]]<|/det|>
+The aim of this case study is to quantify the GHG mitigation potential of different measures or forest management options (in particular afforestation) in terms of meeting an increasing demand for wood, assuming that existing models for the forest sector tend to "underestimate" forest carbon fluxes.  
+
+<|ref|>text<|/ref|><|det|>[[72, 567, 901, 620]]<|/det|>
+In fact, in addition to the development of forest carbon stocks, the delayed release of biogenic carbon through the use of wood as material as well as potential shifting effects of the GHG emission balances associated with the life cycle of these wood- containing product systems and their potential product alternatives do also have an impact on the overall GHG balance.  
+
+<|ref|>text<|/ref|><|det|>[[72, 632, 921, 673]]<|/det|>
+However, the implementation of the presented approach to estimate the total GHG impact of different management scenarios compared to the defined reference, in our view appears to be completely inadequate in this study. It does not comply with applicable international standards and existing state- of- the- art knowledge.  
+
+<|ref|>text<|/ref|><|det|>[[72, 685, 914, 803]]<|/det|>
+While the modeling of carbon storage development in the forest using the internationally recognized Carbon Budget Model for the Canadian Forestry Service appears adequate, the methodological inadequacies relate in particular to the life cycle assessment methodology, on the basis of whose standard- compliant, consistent and transparent implementation avoided emissions through "product substitution" can be estimated in the first place. In order to adequately consider process chain emissions and based on those also "avoided emissions from product substitution", it is crucial to meet the internationally standardized requirements for life cycle assessment (including ISO 14040/44 and ISO 21930). The mere summation of unrelated LCA process information from a background database (here: Ecolnvent) is inadequate - at least for the processes outside the forest along the processing and value chain. The data used for calculating "primary avoided emissions (FF/product substitution)" are also completely unsuitable.  
+
+<|ref|>text<|/ref|><|det|>[[72, 815, 913, 867]]<|/det|>
+Furthermore, the simplified calculation of assumed carbon storage effects through the use of wood as a material (harvested wood products) contradicts central core requirements in the calculation as set out in the methodological guidelines and requirements provided by IPCC (incl. e.g. the consideration of inherited emissions). In consequence, the HWP contribution trough biogenic carbon storage as well as potential "avoided emissions" appear to be massively overestimated.  
+
+<|ref|>text<|/ref|><|det|>[[72, 879, 919, 932]]<|/det|>
+While it seems undisputed, even without the present study, that "the expansion of the (industrial) bioeconomy should be linked to the availability of raw materials in order to avoid unrealistic supply expectations," statements to the effect that "considerable HWP- C storage and product substitution credits can be achieved simultaneously" are not at all tenable on the basis of this simple and, in our view, completely methodologically inadequate implementation.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[70, 46, 923, 75]]<|/det|>
+The simple comparison of a changing supply of forest wood with a modeled demand for this raw material (including potential gaps) can also be carried out without balancing all GHG emissions relevant to the forestry and wood sector.  
+
+<|ref|>text<|/ref|><|det|>[[73, 99, 144, 112]]<|/det|>
+Version 1:  
+
+<|ref|>text<|/ref|><|det|>[[73, 125, 219, 138]]<|/det|>
+Reviewer comments:  
+
+<|ref|>text<|/ref|><|det|>[[73, 151, 160, 164]]<|/det|>
+Reviewer #1  
+
+<|ref|>text<|/ref|><|det|>[[73, 177, 238, 190]]<|/det|>
+(Remarks to the Author)  
+
+<|ref|>text<|/ref|><|det|>[[72, 190, 920, 280]]<|/det|>
+The authors provided fairly comprehensive responses to the reviewers' comments. Yet, after reading the revised paper there is a dissonance between the level of detail in the manuscript and practical applicability of the findings. In the response to reviewers authors noted that this manuscript implements a framework laid out in an earlier paper published in Nature Communications, however the implementation is rather abstract, an interested party (e.g. government of a temperate country) would have to re- do the analyses (i.e. implement the framework to their specific country) and may get substantially different results. This makes me question the value of this particular manuscript, given lack of connection to any specific country, for which the feasibility of the generated estimates could be evaluated.  
+
+<|ref|>text<|/ref|><|det|>[[72, 280, 852, 320]]<|/det|>
+In my opinion, the treatment of the uncertainties associated with the effects of climate change and disturbances on productivity and GWP was fairly simplistic, not illustrated in all figures and not illustrated for GWP estimates. Lastly, please correct the following: CBM- CFS stands for "Carbon Budget Model of the Canadian Forest Sector".  
+
+<|ref|>text<|/ref|><|det|>[[73, 333, 161, 346]]<|/det|>
+Reviewer #2  
+
+<|ref|>text<|/ref|><|det|>[[73, 359, 238, 372]]<|/det|>
+(Remarks to the Author)  
+
+<|ref|>text<|/ref|><|det|>[[72, 740, 916, 793]]<|/det|>
+Open Access This Peer Review File is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.  
+
+<|ref|>text<|/ref|><|det|>[[72, 793, 797, 806]]<|/det|>
+In cases where reviewers are anonymous, credit should be given to 'Anonymous Referee' and the source.  
+
+<|ref|>text<|/ref|><|det|>[[72, 806, 911, 858]]<|/det|>
+The images or other third party material in this Peer Review File are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.  
+
+<|ref|>text<|/ref|><|det|>[[72, 858, 618, 871]]<|/det|>
+To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/
+
+<--- Page Split --->
+<|ref|>title<|/ref|><|det|>[[83, 124, 314, 140]]<|/det|>
+# Author Responses to Reviewer Comments
+
+<|ref|>text<|/ref|><|det|>[[83, 162, 545, 178]]<|/det|>
+We are grateful to and thank the reviewers for taking time to consider our manuscript.
+
+<|ref|>table<|/ref|><|det|>[[81, 195, 911, 810]]<|/det|>
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td></td><td>Reviewer #1:</td><td></td></tr><tr><td></td><td>The manuscript "Can temperate forests deliver both<br>future wood demand and climate-change mitigation?" by Forster et al. presents the results of the lifecycle<br>assessments of the global warming potential of wood<br>supply under the growing wood demand scenarios. The analysis is fairly clear and thorough, however I have<br>several reservations outlined below:</td><td>We thank the reviewer for the positive comments and valuable specific suggestions to<br>further enhance the paper.</td></tr><tr><td>1</td><td>There are no uncertainty estimates around the wood<br>supply and demand as well as GWP. Incorporating the<br>uncertainty analysis into the study would provide<br>insight into where the future research efforts should be directed in order to reduce the uncertainty, as well as<br>help evaluate the confidence in the reported estimates. For an example of uncertainty analysis using CBM-CFS<br>please refer to J.M. Metsaranta, C.H. Shaw, W.A. Kurz,<br>C. Boisvenue, and S. Morken. 2017. Uncertainty of<br>inventory-based estimates of the carbon dynamics of<br>Canada's managed forest (1990-2014). Canadian<br>Journal of Forest Research. 47(8): 1082-1094.<br>https://doi.org/10.1139/cjfr-2017-0088</td><td>The demand profiles modelled are based on the outcome of an extensive literature<br>review of wood demand modelling studies. We recognise and account for the significant<br>uncertainty sourrounding future wood demand projections (indicated by the wide range<br>of demand projections published in the literature) by modelling two different demand<br>profiles in the present study. These two demand profiles capture the range in the<br>published literature. LCA of forest management scenarios under each of these demand<br>profile senarios indicates the degree of uncertainty and sensitivity of results to future<br>demand projection assumptions. In dealing with uncertainties for the supply side, we<br>have taken the simple and clear approach of testing multiple scenarios of variation in<br>future forest productivity (yield). We believe that this is an efficient way of integrating all of these causes of uncertainty listed by the reviewer 1 comment 2. We have also added<br>the following text to the 'Discussion section' lines 375-379: "There is an urgent need for<br>more integrated evidence that incorporates holistic assessment of prospective forestry<br>value chains alongside landscape dynamics (including forest management and<br>expansion), at both national and global scales, including improved estimates of the<br>potential impacts of climate change-linked threats to the future productivity of both<br>temperate forests and the other forest biomes providing wood production."<br>See continuation of this response in response to reviewer 1 comment 2 below.</td></tr></table>
+
+<--- Page Split --->
+<|ref|>table<|/ref|><|det|>[[80, 113, 911, 870]]<|/det|>
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td>2</td><td>Although authors did state in the methods section that risks of pests disease, wind, fire, warming and other climate change effects would not significantly alter the study findings because the risks are highly uncertain and would apply similarly across the study, I don't think it is the case. Fire, drought, changing productivity due to climate change would not change the wood demand, however they could profoundly affect the wood supply. Given the length of the projections in the study, warming would substantially increase heterotrophic respiration, and therefore would alter GWP projections. Drought events could substantially and repeatedly reduce forest yield due via inhibition of the photosynthetic rate and increasing mortality rate. Warming temperatures and changing precipitation regimes are also likely to affect forest net primary production. Increasing fire frequency and severity could profoundly affect the wood supply. I don't think these effects are negligible for GWP and wood supply estimates, and therefore should be considered in the study, especially given the 100-year projection time.</td><td>As described in response to the previous comment, we believe that the approach to test multiple scenarios of variation in future forest productivity (yield) is an efficient way of integrating many of the causes of uncertainty listed by the reviewer. It is the approach that is most compatible with the LCA methodology of our study. Therefore, to address this specific reviewer requirement we have introduced a further pair of scenarios ('lower productivity' than the reference scenario) to model impact of possible ecosystem shocks (leading to yield reduction) that could arise from natural disturbance such as pests, disease, wind, drought and fire. These new scenarios are a modification of the 'reference rotation' existing forest (YC18, 50-yr rotation) + afforestation combination. The new scenarios are a transition of a) 15% and b) 30% of the existing forest from YC18 to YC12 ramping up over a 15 year period (remaining as a 50-yr harvest rotation), followed by recovery back to YC18 after one harvest rotation, ramping back down over a 15-yr period. (The afforestation assumptions were unchanged). Essentially the scenario represents a rapid shock of -two different intensities (affect on 15% &amp;amp; 30% of the total area), followed by recovery after harvest (at 50-yrs old). A detailed description of the basis for the parameterisation of these two scenarios (two and a half pages of text) based on a substantial new literature review (32 references cited) is provided in the new Supplementary Information file "Modelling of natural disturbances - 'reduced productivity'". We recognise that natural disturbances, the intensity and rate of impact, and management responses to these are complex and therefore the modelling options available are vast. We believe the new scenarios selected correspond to contrasting realistic disturbance patterns presented in the reviewed literature and therefore offer additional insights to the other scenarios already modelled. We have included the new results in an updated version of 'Supplementary Data 3', alongside the results of all the other scenarios modelled in this paper.<br>We modelled these two new scenarios under high and low demand projections and for the variation in afforestation rates and planting periods. We found that the new 'lower productivity' scenarios in this set led to higher net forest carbon loss (domestic and overseas) than the 'reference rotation' scenarios, as well as other changes, yet these disturbances do not change the overall results or findings of the study. We have added text on this to the methodology (lines 701-708) and results section of the manuscript</td></tr></table>
+
+<--- Page Split --->
+<|ref|>table<|/ref|><|det|>[[81, 117, 912, 702]]<|/det|>
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses<br>(lines 114-118; 169-172; and 201-204) and have included a more comprehensive description in the new Supplementary Information file (Supplementary Methods 1) in order to minimise additional word count.</td></tr></table>
+
+<--- Page Split --->
+<|ref|>table<|/ref|><|det|>[[80, 115, 912, 530]]<|/det|>
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td>3</td><td>The study results are generalized for a “temperate country”, however most of the data used in the study are for the UK, why not make the study focused on the UK? It would make it easier to assess the feasibility of the proposed scenarios and allow to avoid generalized statements (as a reader, I had a little trouble with those).</td><td>We intentionally kept the analysis generic for a hypothetical temperate country rather than a UK-specific case study in order to provide evidence that has broad relevance. (The UK is not 'typical' in that it has a high baseline of imports and low domestic wood production.) This study is prospective in nature and focuses on the impacts of change, based on transparent assumptions for future changes (e.g. rate of decarbonisation, projected wood demand increase, shifts in forest management, different afforestation rates) that are not unique to the UK (or any other country) or dependant on historic conditions that may be unique to the UK.<br>In order to perform the LCA we needed to make assumptions on typical product breakouts at the forest gate and at a sawmill and a representative wood product flow. Given limitations on public availability of this kind of data, we used data available for the UK as the basis for these assumptions. These assumptions are broadly representative across temperate countries beyond the UK given the similarities in technology and wood value chains. For example, typical rates of conversion of softwood to sawtimber at sawmills is similar across temperate regions. We have edited the Methodology 'Scope of LCA' section (e.g. lines 622-628) to try and improve the reconciliation between our use of UK data and generalised statements.</td></tr></table>
+
+<--- Page Split --->
+<|ref|>table<|/ref|><|det|>[[80, 153, 905, 870]]<|/det|>
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td></td><td>Reviewer #2</td><td></td></tr><tr><td></td><td>Summary: In this manuscript, authors present a combined application of forest carbon models and life cycle assessment (LCA) to estimate global warming (GW) impacts (+/-) of harvested wood products' (HWPs) value chain in a temperate country (appears to be UK). For projected high and low wood demands, the GW impacts are assessed for different<br>scenarios within the country (changing rotation length of forests, increasing rate of production, and expansion of forest area) and overseas imports from non-temperate forests (tropical country). This kind of study is an interesting attempt to integrate two areas of the forest products value chain (forestry and HWPs) for policy implications. The results indicate that increased wood use is not a climate-change solution unless afforestation, increasing forest productivity under sustainable forest management, and mitigating<br>demand increases through enhanced circularity and<br>cascading of wood use are also integrated into the strategy. I am offering below some comments/suggestions to improve the manuscript.</td><td>We thank the reviewer for the positive comments and endorsement of the important findings of the paper.</td></tr><tr><td></td><td>Major comments</td><td></td></tr><tr><td>1</td><td>103-104: The current production and consumption levels of this temperate country should have been characterized to visualize the gap between demand and supply for the<br>reference year 2023, and the how different scenarios or<br>intended decisions might close this gap and influence GW<br>impacts.</td><td>We intentionally defined a hypothetical scenario in which domestic production equals domestic consumption in the reference year 2023, as indicated in Figure 1. As described in our response to reviewer 1 comment 3 above, we have taken this hypothetical approach rather than characterising the actual production and<br>consumption levels of a specific country in order to provide evidence that has<br>broad relevance. (For example, the UK is not 'typical' in that it has a high baseline of imports and low domestic wood production.) It makes sense for a<br>counterfactual lifecycle assessment (which is concerned with change from a<br>baseline) to start the scenario with production = consumption rather that starting</td></tr></table>
+
+<--- Page Split --->
+<|ref|>table<|/ref|><|det|>[[80, 115, 905, 874]]<|/det|>
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td></td><td></td><td>with a deficit or surplus, which would make interpretation of results more challenging.<br>We then visualise how the different scenarios close the gap between future (increasing) demand and supply in Figure 1 by colour coding the supply deficit in red (and the supply surplus in pink). We have also added some additional text in the Figure 1 title to improve this description.</td></tr><tr><td>2</td><td>107-108: What was the reason for selection of Sitka spruce forest in afforestation? The authors might discuss whether the results would be different if the forests were Douglas-fir or Western hemlock.</td><td>Sitka spruce was selected as it grows widely (native to west coast of Canada and the United States; now planted in 16 countries worldwide, including as the predominant plantation species in the UK, Ireland and Denmark - all countries with low forest cover). Sitka spruce also grows well in degraded upland sites where land is most economically viable for afforestation, unlike productive Douglas fir or western hemlock, which require lower altitude land that is of higher value for food production. LCA results would not be different if different species (or species mixes) were modelled unless a significantly different yield class (growth curve) was also assumed. It is the yield rather than the species choice that drives the GWP impact results (which of course would be related in specific geographic contexts but in a hypothetical forest study such as this, the yield class is specified independently by the modeller as a representative average across contexts). In this way, the choice of Sitka spruce is simply a proxy for commercial conifer species. However, we believe that the clarity of the manuscript benefits from naming the case species, which links to the UK sawmill and-representative wood product flow data used in the study.</td></tr><tr><td>3</td><td>Figure 2: Please change y-axis units to Tg CO2e.</td><td>We have made this change.</td></tr><tr><td>4</td><td>234: More interpretations could be added in 'high wood demand projection' results.</td><td>We are pleased that Reviewer 2 sees further possible interpretations that we could include. We would welcome the opportunity to add more interpretations, however because of the journal's strict word limits this would necessitate making significant cuts to other parts of the text, which would sacrifice their clarity and content. Given the importance of these other components of the paper that would have to be cut, with regret we have concluded that adding more</td></tr></table>
+
+<--- Page Split --->
+<|ref|>table<|/ref|><|det|>[[80, 115, 905, 850]]<|/det|>
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td></td><td></td><td>interpretations to the 'high wood demand projection' results would be of net cost to the value of the paper.</td></tr><tr><td>5</td><td>264-267: Not clearly explained how GWP impact of alternative increased due to increase in overseas wood supply. Prolonged use of non-wood product and fuel alternatives?</td><td>We have added words in this sentence (line 290) to clarify that the GWP impact of the alternative is "curtailed HWP supply" i.e. prolonged use of non-wood product and fuel alternatives. We have also added further description to indicate the relative impact of the specific overseas scenarios ('Boreal 1,2&3, Tropical CVL and Tropical (afforestation)' in Fig. 4) in which emissions associated with supplying the wood shortfall are higher or lower that the emissions from 'curtailed HWP supply' (prolonged use of non-wood and fuel alternatives).</td></tr><tr><td>6</td><td>278-279: Please add explanation to this sentence, it seems confusing that more imports from tropical afforestation is better.</td><td>It is unclear to us why this is confusing as it seems to be clearly shown in the results reported in Fig. 4. We do acknowledge the potential disbenefits of new tropical afforestation in lines 363-364 ("numerous socio-economic51 and biodiversity conservation20 caveats". In essence, increasing demand for fast-growing tropical tree species, which could be established on the large areas of degraded land in the tropics (if local socio-economic conditions make them available), could actually enhance terrestrial carbon stocks.</td></tr><tr><td>7</td><td>295: Clarity is needed on what type of non-wood product is considered substituted by HWPs. Because substitution credits for two type of non-wood products can be different for same HWP.</td><td>In response to this point, to avoid adding significant additional word count, we have added text to the methodology (line 616-617) and in the Fig. 5 title to direct readers to supplementary information for more detail on the substitution credits assumptions (supplementary figure 1 and supplementary Table 1). We also direct readers to Forster et al. (2021) where the methodology was originally published. We have also improved figure 5 in the present paper to convey, among other things, the product substitution assumptions more clearly.</td></tr><tr><td>8</td><td>340: BECCS is associated with 'energy substitution' like non-wood products avoided is associated with 'product substitution'. There aren't a lot of interpretations in the results and discussion that focused on energy substitution.</td><td>We have expanded the sentence referring to BECCS to mention product substitution and emphasise the permanent geological storage of biogenic carbon that contributes a carbon sink. Lines 368-369.</td></tr></table>
+
+<--- Page Split --->
+<|ref|>table<|/ref|><|det|>[[80, 113, 905, 860]]<|/det|>
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td>9</td><td>600: there is no ‘ix’ in the components of LCA system boundary. Including via ...?</td><td>We have checked the numbering of components of the LCA system boundary in figure 5 and in the text and cannot find anything missing.</td></tr><tr><td>10</td><td>724-740: This whole paragraph does not seem to be fit for methodology. Its more suitable for introduction.</td><td>We believe that the paragraph is important for justifying the important selection of assumptions for the projected rate of demand increase used in the study, which is a core aspect of the study's methodology. We have edited and changed the order of the text to make the inclusion of this text in the methodology more fluent and justified. E.g. lines 779-785 have been moved from earlier in the section.</td></tr><tr><td>11</td><td>747-748: What is the per capita timber consumption in UK, which translates to 30% increase in demand by 2050?</td><td>We do not believe that including per capita timber consumption in the UK is relevant to the study. The 30% increase by 2050 statistic was included as this is a date that many published projections focus on, because it is an important date for many 'net zero' targets as mandated by the 2015 Paris Agreement of UNFCCC, and therefore a statistic that readers are likely to be interested in. As stated in our response to reviewer 2 comment 1, we have prioritised assumptions that are non-country-specific and forward looking wherever possible, to maximise the transferability of results. In essence, the critical (and transferable) mathematical/biophysical dynamic is the rate of demand increase from the baseline relative to the expansion rate of baseline forest area (through afforestation) and the (change in) productivity of existing (and new) forest.</td></tr><tr><td>12</td><td>754: In this paragraph, it would be good if authors give a brief about the different HWP end-uses (primary and cascading uses) considered and maybe a justification for selecting HWP uses.</td><td>The primary HWP end-uses are described in the methodology and further in the SI, including in a revised system diagram (Figure 5) indicating major processes and products. Details of cascading uses and also justification for the assumptions used are provided in Forster et al. (2021), which we cite in the methodology. We believe there is insufficient justification to repeat all of this detail here, especially on account of the word count restrictions.</td></tr><tr><td>13</td><td>766: More clarity on defining the overseas forest type and carbon storage. Also, was the transportation distance and mode of transport included in the analysis?</td><td>Description of the overseas forest types is provided in the following paragraphs (lines 827-856).<br>We did not include assumed differences in transport distances between overseas and domestic HWPs because, in a previous study (Forster et al., 2023), we found that these differences made a very small contribution (less than 0.01%) relative to the net GWP impact of the value chain.</td></tr></table>
+
+<--- Page Split --->
+<|ref|>table<|/ref|><|det|>[[80, 115, 905, 603]]<|/det|>
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td>14</td><td>785: Why 'GWP (forest C) impact...'? Forest carbon can be stored, emitted, or removed but cannot be equated directly to GWP.</td><td>GWP (forest C) impact is defined in the previous paragraph as the 'GWP impact of fluxes in forest carbon stocks'. We have added '(GWP (forest C))' after this definition to add clarity. It refers to the GWP impact associated with net CO2e fluxes to/from the forest. Line 828.</td></tr><tr><td>15</td><td>Overall, the methods section appears weak to me and needs a thorough revision to ensure that work can be reproduced.</td><td>We have carefully addressed each of the reviewers specific comments and assume that in doing so we have met the need expressed here.</td></tr><tr><td></td><td>Minor comments</td><td></td></tr><tr><td>16</td><td>325 and 330: should it be trip or tip?</td><td>Tip means to tilt, tumble or topple. In this context, 'the balance... can tip from sink to source', means the balance can tilt or shift from sink to source.</td></tr><tr><td>17</td><td>723: why question marks in middle of the sentence?</td><td>We have removed these.</td></tr><tr><td>18</td><td>Double numbering in references</td><td>We have removed these.</td></tr><tr><td>19</td><td>Some of the links in supplementary excel file are broken. Please check.</td><td>We have checked the supplementary Excel files and assume that the reviewer is referring to the 'HWP calculation module' (Supplementary Data 2). The broken links are the emissions factors cells that link to the original LCA worksheet (Supplementary Data 1) that generated them. These links do not work once the workbook is resaved in a different location. However, in the final version of these supplementary data files, which will be stored in a permanent repository or in the publisher's website, we will endeavour to ensure the links all work.</td></tr><tr><td></td><td></td><td></td></tr></table>
+
+<|ref|>table<|/ref|><|det|>[[81, 637, 905, 830]]<|/det|>
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td></td><td>Reviewer #3:</td><td></td></tr><tr><td>1</td><td>The aim of this case study is to quantify the GHG mitigation potential of different measures or forest management<br>options (in particular afforestation) in terms of meeting an<br>increasing demand for wood, assuming that existing models for the forest sector tend to "underestimate" forest carbon fluxes.</td><td></td></tr></table>
+
+<--- Page Split --->
+<|ref|>table<|/ref|><|det|>[[80, 111, 905, 848]]<|/det|>
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td>2</td><td>In fact, in addition to the development of forest carbon stocks, the delayed release of biogenic carbon through the use of wood as material as well as potential shifting effects of the GHG emission balances associated with the life cycle of these wood-containing product systems and their potential product alternatives do also have an impact on the overall GHG balance.</td><td>We are not clear on whether this forms part of the reviewer's summary of the paper or is a comment requiring our response. In case it is the latter, we provide the following clarification. We agree with the reviewer about the importance of impacts throughout the wood products life cycle. Yet, the full climate mitigation effects of forestry are often under-represented because: (i) LCA approaches that (sometimes) better represent substitution and possible end-of-life carbon storage effects are usually only applied to specific products (partial wood flows out of forests); (ii) inventory approaches that attempt to represent all wood products, typically neglect short-lived wood products used for energy, and don't attribute substitution effects back to the forest sector; (iii) both LCA and inventory approaches often disregard cascading uses of wood, and associated second (and possible further) substitutions, along with extended carbon storage effects; (iv) only very prospective analyses consider future BECCS deployment that could lock up biogenic carbon indefinitely.</td></tr><tr><td>3</td><td>However, the implementation of the presented approach to estimate the total GHG impact of different management scenarios compared to the defined reference, in our view appears to be completely inadequate in this study. It does not comply with applicable international standards and existing state-of-the-art knowledge.</td><td>We are surprised by this comment, which we strongly believe does not reflect the content and rigour of our study or its compliance with the highest international standards. International standards apply to specific types of accounting, such as Environmental Product Declarations for (LCA of) wood products or UNFCCC guidelines for national GHG inventory accounting. In contrast, there are no such "standards" for the prospective consequential LCA of entire forest and multiple downstream product systems, which is the innovative methodological approach of our study. That is precisely what makes this approach so valuable - it transcends the accounting rules that are applied to deal with truncated system boundaries (e.g. how to allocate forest effects to specific downstream products). This paper builds on the state-of-the-art approach applied in our earlier, highly-cited manuscript published in Nature Communications (Forster et al., 2021), and applies it to a temperate forest system under different management and expansion regimes; this is highly novel. To reiterate, we are not attempting to calculate product footprints as per various international standards, but instead (as is befitting for a submission to Nature Communications) to provide a novel, holistic and rigorous analysis of the climate mitigation efficacy of different forest</td></tr></table>
+
+<--- Page Split --->
+<|ref|>table<|/ref|><|det|>[[80, 115, 904, 832]]<|/det|>
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td></td><td></td><td>management strategies, providing key new evidence for policy makers and an innovative novel approach for researchers.</td></tr><tr><td>4</td><td>While the modeling of carbon storage development in the forest using the internationally recognized Carbon Budget Model for the Canadian Forestry Service appears adequate, the methodological inadequacies relate in particular to the life cycle assessment methodology, on the basis of whose standard-compliant, consistent and transparent implementation avoided emissions through "product substitution" can be estimated in the first place. In order to adequately consider process chain emissions and based on those also "avoided emissions from product substitution", it is crucial to meet the internationally standardized requirements for life cycle assessment (including ISO 14040/44 and ISO 21930). The mere summation of unrelated LCA process information from a background database (here: Ecolnvent) is inadequate - at least for the processes outside the forest along the processing and value chain. The data used for calculating "primary avoided emissions (FF/product substitution)" are also completely unsuitable.</td><td>As explained in our previous response, our study is categorically not an attributional LCA, which seems to have been the assumption of the reviewer. Instead it is a consequential LCA. That said, this study does fully comply with ISO 14040 and 14044 standards (which in reality are a basic framework to structure LCA), insofar as it: provides a clear goal and scope of the study; calculates an inventory that accounts for relevant changes within the expanded system boundary; applies GWP100 characterisation factors for the life cycle impact assessment phase; interprets results with clear regard for the question and methods used (now including important sensitivity analyses). Furthermore, this study uses system expansion in preference to allocation, as recommended in the allocation hierarchy advocated by the ISO standards. The study categorically does not sum "unrelated" LCA process information - incurred or avoided processes are clearly indicated in the system boundary (Supplementary Fig. 1 and in Fig. 5, also now improved), based on a consequential LCA approach that incorporates all changes associated with the scenarios vs the baseline counterfactual situation.</td></tr></table>
+
+<--- Page Split --->
+<|ref|>table<|/ref|><|det|>[[80, 113, 905, 870]]<|/det|>
+
+<table><tr><td></td><td>Remarks to Author</td><td>Author Responses</td></tr><tr><td>5</td><td>Furthermore, the simplified calculation of assumed carbon storage effects through the use of wood as a material (harvested wood products) contradicts central core requirements in the calculation as set out in the methodological guidelines and requirements provided by IPCC (incl. e.g. the consideration of inherited emissions). In consequence, the HWP contribution trough biogenic carbon storage as well as potential “avoided emissions” appear to be massively overestimated.</td><td>Our methodology does not contradict core requirements of the IPCC and from this comment we do not understand how the reviewer considers that it does so. The IPCC provide guidelines for a number of approaches to deal with HWP C storage at a national scale. All of them are based on the fact that biogenic carbon harvested from the forest is "lost" from the terrestrial system, but some of this carbon "reappears" in longer-lived HWPs (with a large deficit reflecting biomass assumed to be immediately oxidised via combustion for bioenergy or for kiln drying of wood in sawmills). Our approach fully respects the biogenic carbon balance - in fact much more explicitly than typical national inventory accounting (with which we are fully familiar), because flows into all main products, both short- and long-lived, are accounted for, and the duration of carbon storage in each of these products, and any subsequent products, is explicitly accounted for in our methodology. Again, this is the benefit of our expanded boundary approach (across products and through time) - it minimises the influence of "cut-off" rules and associated value judgements, which often profoundly influence carbon footprints at the wood product level.</td></tr><tr><td>6</td><td>While it seems undisputed, even without the present study, that "the expansion of the (industrial) bioeconomy should be linked to the availability of raw materials in order to avoid unrealistic supply expectations, " statements to the effect that "considerable HWP-C storage and product substitution credits can be achieved simultaneously" are not at all tenable on the basis of this simple and, in our view, completely methodologically inadequate implementation.</td><td>Without any further insight from the reviewer as to where our methodology is inadequate, it is difficult to address this comment. We acknowledge that our original system diagram did not fully represent the scope of our methodology (although it was described fully in the text and illustrated in Supplementary Fig.1). Therefore, we hope that the improved system diagram in the resubmitted manuscript helps understanding of precisely how we have linked forest C storage with HWP-C storage and substitution, and long-term CCS C-storage effects in a coherent manner. As previously mentioned, this builds on the methodology of Forster et al. (2021) published in Nature Communications.</td></tr><tr><td>7</td><td>The simple comparison of a changing supply of forest wood with a modeled demand for this raw material (including potential gaps) can also be carried out without balancing all GHG emissions relevant to the forestry and wood sector.</td><td>It is true that a simple comparison of changing supply of forest wood with demand could be carried out in isolation. However, that would not address the important policy question of what the climate mitigation effect would be of alternative forest strategies that aim to minimise the gap between future supply and demand. That is where this paper provides a unique and robust contribution to the scientific literature, for which its rigorous methodological consequential LCA approach is essential.</td></tr></table>
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[0, 0, 997, 997]]<|/det|>
+# 1.1.1.1.1.1.1.1.1.1.1
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[120, 85, 426, 100]]<|/det|>
+## Author response to reviewer comments  
+
+<|ref|>text<|/ref|><|det|>[[119, 111, 848, 162]]<|/det|>
+We are once again grateful for the opportunity to respond to Reviewer comments. We have addressed the comments below and in doing so believe we have further enhanced the article, as outlined in our response. All changes to the manuscript are highlighted in 'tracked- changes'.  
+
+<|ref|>sub_title<|/ref|><|det|>[[119, 199, 298, 214]]<|/det|>
+## REVIEWER COMMENTS  
+
+<|ref|>text<|/ref|><|det|>[[120, 233, 404, 248]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>sub_title<|/ref|><|det|>[[119, 268, 210, 283]]<|/det|>
+## Comment 1  
+
+<|ref|>text<|/ref|><|det|>[[118, 293, 879, 448]]<|/det|>
+The authors provided fairly comprehensive responses to the reviewers' comments. Yet, after reading the revised paper there is a dissonance between the level of detail in the manuscript and practical applicability of the findings. In the response to reviewers authors noted that this manuscript implements a framework laid out in an earlier paper published in Nature Communications, however the implementation is rather abstract, an interested party (e.g. government of a temperate country) would have to re- do the analyses (i.e. implement the framework to their specific country) and may get substantially different results. This makes me question the value of this particular manuscript, given lack of connection to any specific country, for which the feasibility of the generated estimates could be evaluated.  
+
+<|ref|>sub_title<|/ref|><|det|>[[119, 460, 308, 475]]<|/det|>
+## Response to Comment 1  
+
+<|ref|>text<|/ref|><|det|>[[117, 485, 876, 760]]<|/det|>
+We thank Reviewer 1 for their further comments, to which we have given careful consideration. The main point raised here is that the methodology remains abstract, such that others would need to perform their own analysis. We fully agree that researchers and other interested parties should perform context- specific analysis, and not rely solely on the results of a generic country case study (of the kind we use here to demonstrate the rigour, power and applicability of the framework that we have developed in this paper). We have subtly revised framing to reflect this, with additional clarity and emphasis on the novel methodological framework we have developed. In a new SI, we outline the framework in diagram, table and text form, highlighting its substantial novel components not previously published, and describe its implementation with reference to important parameters and examples of datasets and models of the kind that are available to generate context- specific results. However, we maintain that using a generic temperate country with an even- aged forest as an illustrative case study provides important insights that do have broad relevance for readers, whereas using the context of a specific country would not, owing to the unique nature of its forest- age deviations. We have made small edits throughout the manuscript to further emphasise to readers the importance of tailoring studies to specific contexts, especially where the results of such studies are intended to inform policy.  
+
+<|ref|>text<|/ref|><|det|>[[118, 770, 864, 838]]<|/det|>
+The framework sets out our exploratory approach for identifying 'low regrets' climate solutions for the forestry value chain by evaluating a range of plausible future scenarios, using powerful integrated forest modelling and lifecycle assessment (LCA) (Supplementary Fig. 1 & Supplementary Methods 1).  
+
+<|ref|>text<|/ref|><|det|>[[118, 848, 870, 900]]<|/det|>
+The present study builds substantially on the detailed prospective dynamic LCA modelling of entire forest- wood value chains developed and explained in Forster et al. (2021). Crucially, in the present study the full downstream greenhouse gas (GHG) mitigation consequences of wood use and end- of
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[118, 84, 853, 186]]<|/det|>
+life management are considered - including carbon storage and material and energy substitution within a decarbonising future economy. Novel aspects to the present study (shown in pink in the new Supplementary Fig. 1) include calculating potential future wood supply deficit by comparing projected wood demand curves to wood supply from a range of modelled (expanded) temperate forest management scenarios, and linking the supply deficit to marginal expansion of supply from other regions.  
+
+<|ref|>text<|/ref|><|det|>[[118, 196, 857, 248]]<|/det|>
+Elaboration of the framework within the SI will facilitate interested parties to follow our approach for a relevant study context (Supplementary Table 1). We signpost readers to this SI at multiple points in the manuscript.  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 276, 210, 291]]<|/det|>
+## Comment 2  
+
+<|ref|>text<|/ref|><|det|>[[118, 302, 866, 352]]<|/det|>
+In my opinion, the treatment of the uncertainties associated with the effects of climate change and disturbances on productivity and GWP was fairly simplistic, not illustrated in all figures and not illustrated for GWP estimates.  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 364, 308, 379]]<|/det|>
+## Response to Comment 2  
+
+<|ref|>text<|/ref|><|det|>[[117, 390, 875, 544]]<|/det|>
+We acknowledge the simple treatment of climate change and disturbances (which we note can vary considerably, geographically) for our generic temperate country study. However, we maintain that this approach is appropriate to illustrate the magnitude of effect associated with highly uncertain future events. Recent articles pertaining to climate mitigation modelling highlight the need for "robust" rather than optimised decision making in the face of deep uncertainty and high complexity (Workman et al. (2021) and Workman et al. (2024)). We now cite these articles to further justify our approach, and propose natural disturbance events as a key parameter to represent (where possible) this impact within the proposed framework when tailoring scenarios in future studies for specific context.  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 555, 210, 569]]<|/det|>
+## Comment 3  
+
+<|ref|>text<|/ref|><|det|>[[118, 581, 839, 613]]<|/det|>
+Lastly, please correct the following: CBM- CFS stands for "Carbon Budget Model of the Canadian Forest Sector".  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 625, 308, 640]]<|/det|>
+## Response to Comment 3  
+
+<|ref|>text<|/ref|><|det|>[[118, 653, 258, 667]]<|/det|>
+Corrections made.
+
+<--- Page Split --->

peer_reviews/1265f8b06778239434bb4b6e3d1588d191708414dc2de58cfdb2da9350869262/supplementary_0_Peer Review File/images_list.json

@@ -0,0 +1 @@
+[]

peer_reviews/1265f8b06778239434bb4b6e3d1588d191708414dc2de58cfdb2da9350869262/supplementary_0_Peer Review File/supplementary_0_Peer Review File.mmd

@@ -0,0 +1,123 @@
+
+# nature portfolio  
+
+Peer Review File  
+
+Systematic Review and Meta- Analysis for a Global Patient co- Owned Cloud (GPOC)  
+
+![PLACEHOLDER_0_0]
+  
+
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+
+Reviewers' Comments:  
+
+Reviewer #1:  
+
+Remarks to the Author:  
+
+The article systematically reviews sharing of electronic health records in a cloud. The text is original, follows PRISMA protocol, and presents exciting insights to the scientific community. The authors considered the most significant articles on the subject. The protocol is adequately done and gives a discussion pointing out the main challenges and issues.  
+
+Following, I list some suggestions to improve the article:  
+
+- GPOC is not a universally accepts term. It is a term that the authors have used in previous publications. Although I do not see a problem using this term in the article, the authors could better define what it means and why it could not use just PHR to represent this idea. Another doubt is what is a co-owned cloud. Is it a public cloud? Is it co-owned by the patient and its health providers?  
+
+- Pg. 3, line 45: The authors discuss centralized approaches. It would be interesting to mention fog-based and peer-to-peer/hierarchical methods for completeness.  
+
+- Pg. 3, line 63: How to deal with the language barrier among cross-border travel? I would appreciate a discussion on this subject.  
+
+- When discussing security and privacy, one of the focuses of the article, you could analyze the implications of HIPAA and GDPR. There are many aspects that these kinds of legislation affect privacy regarding health records and their relation to cross-border travel.  
+
+- Security is mainly discussed in terms of encryption and decryption of data. A more comprehensive analysis is needed, including the three main aspects of confidentiality, integrity, and availability.  
+
+- The method is well-defined and based on PRISMA. I did not find the range period for the search, the bases used, and the specific terms (on page 5, line 94, you only mention some keywords).  
+
+- Interoperability is an important issue that is not discussed at all. What is the rationale, and why was it not considered in the article? A global trend of adopting HL7 FHIR and other standards could influence the field.  
+
+- Security-based parameters are based on times in ms: time for encryption, time for decryption, and the ratio of means. These times depend on the infrastructure: CPU, memory, network bandwidth, etc. Just showing the number of effective time adds little value since it depends on the infrastructure used.  
+
+- I missed a big picture of the main issues in the discussion section. The discussion is based on the more traditional "Author X studied that; Author Y discusses." You could start with a big picture presenting the main topics and then more implicitly bringing the related studies.  
+
+- You cite types of encryptions on lines 281-282 (page 14). A discussion could be added because it could influence many essential aspects of the study.  
+
+- Blockchain is cited throughout the article, and I consider it a trend. You could discuss this technology further and its impact with more emphasis. Other crucial subjects still need to be explored, including Federated Learning, Fog Computing, and the Internet of Things (how to combine data from wearables from patient EHR, for instance).  
+
+- On page 19 you discuss machine learning and the use of data for decision, preserving privacy, and not disclosing patients' details. Anonymization and obfuscation, among other methods, could be commented on here. Although machine learning is not a focus of the article, it could influence GPOC in the services provided and organization/storage of data.
+
+<--- Page Split --->
+
+- Line 386, Pg. 20, you cite "four main areas," but it seems to be five. It would be nice to see some future directions.  
+
+Reviewer #2:  
+
+Remarks to the Author:  
+
+I appreciate and congratulate all authors for adding knowledge to the Global Patient co- Owned Cloud (GPOC) a Cloud- based infrastructure that shares information in the Personal Health Records (PHRs) space. This topic is timely and requires significant effort to establish particularly the concern on data security and privacy. I believe the meta- analysis and results presented cover the area to some extent however the term "use- ubiquity globally" that appears in the abstract does not take any advantages and/or support to the claims (refer to the conclusion).  
+
+Some areas to improve based on the Electronic Health Records (EHRs) use are:  
+
+(1) Validity and Bias: This is very critical to explain how the risk minimisation was managed.  
+
+(2) Perhaps, the language tome could be lower to support the general audience and/or readers (since this is Open Access Forum);  
+
+(3) In addition, Blockchain protocol-embedded PHRs might be another approach that did not dominate the claim hence expanding such idea and thinking would be an added advantage to the paper.  
+
+(4) The conclusion section:  
+
+I believe there are standards when sharing EHRs. Such standards are depending on the Geographical locations, the Country's data protection, and/or breaches of legislation and their maturity (this is a very challenging area indeed).  
+
+## ADDITIONAL COMMENTS:  
+
+The manuscripts share some form of information from the literature search. While they are interconnected, scientifically it is appropriate and valid to share the information however, the merit of the research contributions is somewhat diluted. Hence, it would be advisable to realign the findings specifically to the papers titled, for the readers' comfort and benefits.  
+
+In general, the body of work completed was commendable. At the same time, the written language may not be palatable to a wider audience and readers. It would be advisable to make it simple where possible since there are several paragraphs confounded with each other.  
+
+One of the classic examples in the paper titled "Systematic Review and Meta- Analysis for a Global Patient co- Owned Cloud (GPOC)" uses the term "use- ubiquity globally" in the abstract. This term is misleading based on the conclusion.
+
+<--- Page Split --->
+
+Please see the below answers and actions to all the 30 points of the major revision.
+
+(1) The manuscript **GPOC SYS-META Clean** does not contain comments, it is a revised clean manuscript.
+
+(2) The manuscript **GPOC SYS-META Comms** - you find the actions labelled with comments #1-30.
+
+(3) The manuscript **GPOC SYS-META Highlighted** - you find all the changes underlined and highlighted.
+
+(4) For convenience, the manuscript **GPOC SYS-META Trace** (Word-tracked) is fully traceable from the previous version that you reviewed.
+
+<table><tr><td>Comments</td><td>#</td><td>Part</td><td>Point by Point Answers & Actions</td></tr><tr><td>Editor comments found in the<br>separate cover letter.</td><td>1-8</td><td>All</td><td>See author cover letter.</td></tr><tr><td colspan="4">REVIEWER COMMENTS</td></tr><tr><td colspan="4">Reviewer #1 (Remarks to the<br>Author)</td></tr><tr><td>The article systematically<br>reviews sharing of electronic<br>health records in a cloud. The<br>text is original, follows<br>PRISMA protocol, and presents<br>exciting insights to the<br>scientific community. The<br>authors considered the most<br>significant articles on the<br>subject. The protocol is<br>adequately done and gives a<br>discussion pointing out the<br>main challenges and issues.</td><td>9</td><td>All</td><td>We thank the reviewer for constructive and helpful comments that has ameliorated the manuscript.</td></tr><tr><td>Following, I list some<br>suggestions to improve the<br>article: GPOC is not a<br>universally accepts term. It is a term that the authors have used<br>in previous publications.<br>Although I do not see a<br>problem using this term in the<br>article, the authors could better<br>define what it means and why it could not use just PHR to<br>represent this idea. Another<br>doubt is what is a co-owned<br>cloud. Is it a public cloud? Is it<br>co-owned by the patient and its<br>health providers?</td><td>10</td><td>Abst.<br>&<br>Intro.</td><td>It is important to define the unicity of the GPOC concept. This has now been implemented into the abstract and introduction. Hence, now there is a better definition of GPOC, what the co-ownership is, and why it is trisected for legal reasons. It has also been highlighted why a just PHR cannot be used - the<br>unicity of the GPOC concept of cloud, ownership, AI<br>integration, blockchain, foundation, globality, sharing,<br>independence, interaction, legal foundation status and<br>substrate for development of ML and its global dissemination has been elaborated.</td></tr><tr><td>- Pg. 3, line 45: The authors<br>discuss centralized approaches.<br>It would be interesting to<br>mention fog-based and peer-to-peer/hierarchical methods for<br>completeness.</td><td>11</td><td>Disc.</td><td>These have now been implemented as valuable new angles.<br>Now the fog-based and peer-to-peer/hierarchical methods are mentioned.</td></tr></table>
+
+<--- Page Split --->
+
+
+<table><tr><td>- Pg. 3, line 63: How to deal with the language barrier among cross-border travel? I would appreciate a discussion on this subject.</td><td>12</td><td>Disc.</td><td>This is now included in the discussion among the challenges and importance of AI integration into GPOC. We exemplify how inbuilt AI in the GPOC can solve language barriers.</td></tr><tr><td>- When discussing security and privacy, one of the focuses of the article, you could analyze the implications of HIPAA and GDPR. There are many aspects that these kinds of legislation affect privacy regarding health records and their relation to cross-border travel.</td><td>13</td><td>Disc.</td><td>Thank you, we have now added a comment about and discuss this. After this comment we refer to the GPOC Ethics article 4 in the reference list. This article focuses entirely on ethics, policy, legal, regulation, etc. This is a self-contained article within the GPOC series. GDPR &amp;amp; HIPAA now mentioned (without infringing into the Ethics' article), but now highlighted. Indeed, we cross reference to the article that focuses in this area.</td></tr><tr><td>- Security is mainly discussed in terms of encryption and decryption of data. A more comprehensive analysis is needed, including the three main aspects of confidentiality, integrity, and availability.</td><td>14</td><td>Disc.</td><td>Now included and analysed articles of the systematic review elaborating these main aspects of confidentiality (19), integrity (11), and availability (12). Within parenthesis the number of articles bringing this up.</td></tr><tr><td>- The method is well-defined and based on PRISMA. I did not find the range period for the search, the bases used, and the specific terms (on page 5, line 94, you only mention some keywords).</td><td>15</td><td>Meth. &amp;amp; Supp.</td><td>The PRISMA form and other search documents have now all been included as supplements and referred to as supplementary files S1 and S3. In S1 the full search strategy is visible, with search strings etc. In S3 the PRISMA checklist is presented.</td></tr><tr><td>- Interoperability is an important issue that is not discussed at all. What is the rationale, and why was it not considered in the article? A global trend of adopting HL7 FHIR and other standards could influence the field.</td><td>16</td><td>Disc.</td><td>Now interoperability has been implemented into the discussion. Now also HL7 FHIR is mentioned. Also, as cross referencing here has been made to the technical EcoTech Article.2</td></tr><tr><td>- Security-based parameters are based on times in ms: time for encryption, time for decryption, and the ratio of means. These times depend on the infrastructure: CPU, memory, network bandwidth, etc. Just showing the number of effective time adds little value since it depends on the infrastructure used.</td><td>17</td><td>Disc.</td><td>Now this point has been added to the big picture of the results, as a decisive factor to consider and weigh in. The infrastructure for GPOC is further elaborated in the technical EcoTech Article.2 It is cross referenced to this now too.</td></tr><tr><td>- I missed a big picture of the main issues in the discussion section. The discussion is based on the more traditional "Author X studied that; Author Y discusses." You could start with a big picture presenting the main topics and then more implicitly bringing the related studies.</td><td>18</td><td>Disc.</td><td>Now the big picture has been added in an initial overview, but also in the discussion. Al the places with the traditionally mention "author X studied Y", etc, have been altered and the result is a much better flow. Now, with an accompanying big picture before and after, it is more readable. As advised the big picture is presented mainly first. Also, the traditional author mentioning has been partly altered where possible now, so that the sentences are more neutral as flow better for the reader.</td></tr><tr><td>- You cite types of encryptions on lines 281-282 (page 14). A</td><td>19</td><td>Disc.</td><td>A discussion has now been added on encryption. Also, referring further the technical EcoTech Article.2</td></tr></table>
+
+<--- Page Split --->
+
+
+<table><tr><td>discussion could be added<br>because it could influence many<br>essential aspects of the study.</td><td></td><td></td><td></td></tr><tr><td>- Blockchain is cited throughout the article, and I consider it a<br>trend. You could discuss this<br>technology further and its<br>impact with more emphasis.<br>Other crucial subjects still need<br>to be explored, including<br>Federated Learning, Fog<br>Computing, and the Internet of<br>Things (how to combine data<br>from wearables from patient<br>EHR, for instance).</td><td>20</td><td>Disc.</td><td>Now a discussion on blockchain has been added. Federated Learning, Fog Computing, and the Internet of Things are mentioned, but also a refence for further reading of the technical EcoTech Article.2</td></tr><tr><td>- On page 19 you discuss<br>machine learning and the use of<br>data for decision, preserving<br>privacy, and not disclosing<br>patients' details. Anonymization and obfuscation, among other<br>methods, could be commented<br>on here. Although machine<br>learning is not a focus of the<br>article, it could influence GPOC in the services provided and<br>organization/storage of data.</td><td>21</td><td>Disc.</td><td>Now anonymization and obfuscation and also fully<br>homomorphic encryption is mentioned. There is now also a refence for further reading, where this is elaborated, in the technical EcoTech Article.2</td></tr><tr><td>- Line 386, Pg. 20, you cite<br>"four main areas," but it seems<br>to be five. It would be nice to<br>see some future directions.</td><td>22</td><td>Disc.</td><td>It has now been changed to eight (8) areas. Moreover, the initial keywords of each numbered paragraph have been italicised for increased clarity and readability. Also, the<br>number of directions has been expanded and are now also synched with the six problems statements mentioned before and also referred to in the GPOC series. Again, without any overlap.</td></tr><tr><td>Reviewer #2 (Remarks to the<br>Author)</td><td></td><td></td><td></td></tr><tr><td>I appreciate and congratulate all<br>authors for adding knowledge<br>to the Global Patient co-Owned Cloud (GPOC) a Cloud-based<br>infrastructure that shares<br>information in the Personal<br>Health Records (PHRs) space.<br>This topic is timely and requires significant effort to establish<br>particularly the concern on data security and privacy. I believe<br>the meta-analysis and results<br>presented cover the area to<br>some extent however the term<br>"use-ubiquity globally" that<br>appears in the abstract does not<br>take any advantages and/or<br>support to the claims (refer to<br>the conclusion).</td><td>23</td><td>Abst.</td><td>Many thanks. And yes, the term "use-ubiquity globally" that appears in the abstract, has been deleted and the whole<br>abstract has been completely rewritten. It is now within the<br>required limits of &lt;150 words. Decreased from 244 words.<br>The abstract now starts to explain GPOC, present the series to get a context for the reader, and then focuses on the article and what it shows.</td></tr><tr><td>Some areas to improve based<br>on the Electronic Health<br>Records (EHRs) use are: 1)<br>Validity and Bias: This is very</td><td>24</td><td>Res.</td><td>Now under the headline "Validity and Bias" it has now been<br>explained how the risk minimisation was managed.</td></tr></table>
+
+<--- Page Split --->
+
+
+<table><tr><td>critical to explain how the risk minimisation was managed.</td><td></td><td></td><td></td></tr><tr><td>(2) Perhaps, the language tone could be lower to support the general audience and/or readers (since this is Open Access Forum)</td><td>25</td><td>All</td><td>As the editor also requested (and referred to reviewer 2) above in comment #1, very sentence of the whole manuscript has been put under revision and all paragraphs and nearly all sentences have been altered or polished. Hence, the language is now more suitable to a general audience and the sentences are shorter.</td></tr><tr><td>(3) In addition, Blockchain protocol-embedded PHRs might be another approach that did not dominate the claim hence expanding such idea and thinking would be an added advantage to the paper.</td><td>26</td><td>Disc.</td><td>Now the approach with Blockchain protocol-embedded PHRs have been mentioned. Reference added for further reading of the technical EcoTech Article.²</td></tr><tr><td>(4) The conclusion section: I believe there are standards when sharing EHRs. Such standards are depending on the Geographical locations, the Country's data protection, and/or breaches of legislation and their maturity (this is a very challenging area indeed).</td><td>27</td><td>Disc. & Con.</td><td>Now a hint about this is added. This is the full subject and ranger of the Ethics Article⁴. This article fully delves into geographical variations of ethics, legislation, regulations and policies. It is an additional combinatory literature review and interview series on this subject. It could not have been included into the present systematic review and meta-analysis and is self-contained and stands on its own legs. But a cross referencing has been added with the mentioned hint that this is important, and worth an independent article. See also the last point 8 of the discussion summary.</td></tr><tr><td colspan="4">ADDITIONAL COMMENTS</td></tr><tr><td>The manuscripts share some form of information from the literature search. While they are interconnected, scientifically it is appropriate and valid to share the information however, the merit of the research contributions is somewhat diluted. Hence, it would be advisable to realign the findings specifically to the papers titled, for the readers' comfort and benefits.</td><td>28</td><td>Abst. & Intro</td><td>Absolutely. The overlaps have now been minimised. Hopefully eradicated. For instance, it is requested in the revision to delve into a technical or regulatory area, and these have now been added, but only very brief and with a reference to the article which has this area as its main subject. It is hence important that the required technical additions to the systematic review do not infringe or dilute other manuscripts. Note: all mentioning of subjects belonging to another manuscript has been deleted. Similarly other subjects belonging to other self-contained parts of the series have been deleted and referred to only once for information to the reader (ref 1,2,3,4). Hence, a total realignment has been carried out.</td></tr><tr><td>In general, the body of work completed was commendable. At the same time, the written language may not be palatable to a wider audience and readers. It would be advisable to make it simple where possible since there are several paragraphs confounded with each other.</td><td>29</td><td>All</td><td>The whole text has now been rewritten. The language has been improved and the text flows better. It is adapted to the general audience. The sentences are less complex and shorter. Easier terminology has been used, whenever possible. When comparing the previous version and the present article manuscript all paragraphs and a majority of sentences have been rewritten to ease the flow for the general audience, which is very important. The paragraphing has been changed to eradicate any confusion. Note that now both the abstract (in two short sentences) and the introduction's first paragraph briefly mention the GPOC concept and the context of the GPOC series. Hence, the reader will immediately know what to expect from the articles regarding scope and contents. At the onset the reader will be aware that the five GPOC series entities will display distinct and non-overlapping foci. Hence, this sets the tone and contributes to the realignment of the articles. Several paragraphs hence became superfluous and have been deleted.</td></tr></table>
+
+<--- Page Split --->
+
+
+<table><tr><td>One of the classic examples in the paper titled "Systematic Review and Meta-Analysis for a Global Patient co-Owned Cloud (GPOC)" uses the term "use-ubiquity globally" in the abstract. This term is misleading based on the conclusion.</td><td>30</td><td>Abst.</td><td>This term has been deleted. The abstract has been completely rewritten. It has also been shortened from 244 to the required 150 words maximum. All intricate sentences with complex syntax and prosody have been altered. The message is clearer. Sentences shorter. Language simpler.</td></tr></table>
+
+<--- Page Split --->
+
+Reviewers' Comments:  
+
+Reviewer #1:  
+
+Remarks to the Author:  
+
+The authors have conducted all the modifications asked for in the previous review. The article is now ready to be accepted.
+
+<--- Page Split --->

peer_reviews/1265f8b06778239434bb4b6e3d1588d191708414dc2de58cfdb2da9350869262/supplementary_0_Peer Review File/supplementary_0_Peer Review File_det.mmd

@@ -0,0 +1,163 @@
+<|ref|>title<|/ref|><|det|>[[61, 40, 508, 90]]<|/det|>
+# nature portfolio  
+
+<|ref|>text<|/ref|><|det|>[[70, 110, 362, 139]]<|/det|>
+Peer Review File  
+
+<|ref|>text<|/ref|><|det|>[[70, 154, 912, 211]]<|/det|>
+Systematic Review and Meta- Analysis for a Global Patient co- Owned Cloud (GPOC)  
+
+<|ref|>image<|/ref|><|det|>[[57, 732, 240, 782]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[250, 732, 911, 785]]<|/det|>
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[119, 84, 293, 97]]<|/det|>
+Reviewers' Comments:  
+
+<|ref|>text<|/ref|><|det|>[[119, 112, 223, 125]]<|/det|>
+Reviewer #1:  
+
+<|ref|>text<|/ref|><|det|>[[119, 127, 300, 140]]<|/det|>
+Remarks to the Author:  
+
+<|ref|>text<|/ref|><|det|>[[119, 140, 860, 197]]<|/det|>
+The article systematically reviews sharing of electronic health records in a cloud. The text is original, follows PRISMA protocol, and presents exciting insights to the scientific community. The authors considered the most significant articles on the subject. The protocol is adequately done and gives a discussion pointing out the main challenges and issues.  
+
+<|ref|>text<|/ref|><|det|>[[119, 210, 551, 224]]<|/det|>
+Following, I list some suggestions to improve the article:  
+
+<|ref|>text<|/ref|><|det|>[[119, 238, 874, 309]]<|/det|>
+- GPOC is not a universally accepts term. It is a term that the authors have used in previous publications. Although I do not see a problem using this term in the article, the authors could better define what it means and why it could not use just PHR to represent this idea. Another doubt is what is a co-owned cloud. Is it a public cloud? Is it co-owned by the patient and its health providers?  
+
+<|ref|>text<|/ref|><|det|>[[118, 322, 848, 351]]<|/det|>
+- Pg. 3, line 45: The authors discuss centralized approaches. It would be interesting to mention fog-based and peer-to-peer/hierarchical methods for completeness.  
+
+<|ref|>text<|/ref|><|det|>[[118, 364, 812, 393]]<|/det|>
+- Pg. 3, line 63: How to deal with the language barrier among cross-border travel? I would appreciate a discussion on this subject.  
+
+<|ref|>text<|/ref|><|det|>[[119, 406, 847, 448]]<|/det|>
+- When discussing security and privacy, one of the focuses of the article, you could analyze the implications of HIPAA and GDPR. There are many aspects that these kinds of legislation affect privacy regarding health records and their relation to cross-border travel.  
+
+<|ref|>text<|/ref|><|det|>[[119, 461, 848, 504]]<|/det|>
+- Security is mainly discussed in terms of encryption and decryption of data. A more comprehensive analysis is needed, including the three main aspects of confidentiality, integrity, and availability.  
+
+<|ref|>text<|/ref|><|det|>[[118, 517, 870, 547]]<|/det|>
+- The method is well-defined and based on PRISMA. I did not find the range period for the search, the bases used, and the specific terms (on page 5, line 94, you only mention some keywords).  
+
+<|ref|>text<|/ref|><|det|>[[118, 560, 875, 602]]<|/det|>
+- Interoperability is an important issue that is not discussed at all. What is the rationale, and why was it not considered in the article? A global trend of adopting HL7 FHIR and other standards could influence the field.  
+
+<|ref|>text<|/ref|><|det|>[[119, 616, 874, 672]]<|/det|>
+- Security-based parameters are based on times in ms: time for encryption, time for decryption, and the ratio of means. These times depend on the infrastructure: CPU, memory, network bandwidth, etc. Just showing the number of effective time adds little value since it depends on the infrastructure used.  
+
+<|ref|>text<|/ref|><|det|>[[118, 686, 875, 728]]<|/det|>
+- I missed a big picture of the main issues in the discussion section. The discussion is based on the more traditional "Author X studied that; Author Y discusses." You could start with a big picture presenting the main topics and then more implicitly bringing the related studies.  
+
+<|ref|>text<|/ref|><|det|>[[118, 742, 875, 771]]<|/det|>
+- You cite types of encryptions on lines 281-282 (page 14). A discussion could be added because it could influence many essential aspects of the study.  
+
+<|ref|>text<|/ref|><|det|>[[119, 785, 830, 841]]<|/det|>
+- Blockchain is cited throughout the article, and I consider it a trend. You could discuss this technology further and its impact with more emphasis. Other crucial subjects still need to be explored, including Federated Learning, Fog Computing, and the Internet of Things (how to combine data from wearables from patient EHR, for instance).  
+
+<|ref|>text<|/ref|><|det|>[[119, 855, 863, 911]]<|/det|>
+- On page 19 you discuss machine learning and the use of data for decision, preserving privacy, and not disclosing patients' details. Anonymization and obfuscation, among other methods, could be commented on here. Although machine learning is not a focus of the article, it could influence GPOC in the services provided and organization/storage of data.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[118, 97, 835, 125]]<|/det|>
+- Line 386, Pg. 20, you cite "four main areas," but it seems to be five. It would be nice to see some future directions.  
+
+<|ref|>text<|/ref|><|det|>[[118, 167, 222, 180]]<|/det|>
+Reviewer #2:  
+
+<|ref|>text<|/ref|><|det|>[[118, 181, 298, 194]]<|/det|>
+Remarks to the Author:  
+
+<|ref|>text<|/ref|><|det|>[[118, 195, 870, 281]]<|/det|>
+I appreciate and congratulate all authors for adding knowledge to the Global Patient co- Owned Cloud (GPOC) a Cloud- based infrastructure that shares information in the Personal Health Records (PHRs) space. This topic is timely and requires significant effort to establish particularly the concern on data security and privacy. I believe the meta- analysis and results presented cover the area to some extent however the term "use- ubiquity globally" that appears in the abstract does not take any advantages and/or support to the claims (refer to the conclusion).  
+
+<|ref|>text<|/ref|><|det|>[[118, 293, 725, 308]]<|/det|>
+Some areas to improve based on the Electronic Health Records (EHRs) use are:  
+
+<|ref|>text<|/ref|><|det|>[[118, 308, 660, 336]]<|/det|>
+(1) Validity and Bias: This is very critical to explain how the risk minimisation was managed.  
+
+<|ref|>text<|/ref|><|det|>[[118, 363, 848, 392]]<|/det|>
+(2) Perhaps, the language tome could be lower to support the general audience and/or readers (since this is Open Access Forum);  
+
+<|ref|>text<|/ref|><|det|>[[118, 404, 864, 447]]<|/det|>
+(3) In addition, Blockchain protocol-embedded PHRs might be another approach that did not dominate the claim hence expanding such idea and thinking would be an added advantage to the paper.  
+
+<|ref|>text<|/ref|><|det|>[[118, 461, 325, 475]]<|/det|>
+(4) The conclusion section:  
+
+<|ref|>text<|/ref|><|det|>[[118, 476, 837, 519]]<|/det|>
+I believe there are standards when sharing EHRs. Such standards are depending on the Geographical locations, the Country's data protection, and/or breaches of legislation and their maturity (this is a very challenging area indeed).  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 533, 316, 546]]<|/det|>
+## ADDITIONAL COMMENTS:  
+
+<|ref|>text<|/ref|><|det|>[[118, 547, 876, 603]]<|/det|>
+The manuscripts share some form of information from the literature search. While they are interconnected, scientifically it is appropriate and valid to share the information however, the merit of the research contributions is somewhat diluted. Hence, it would be advisable to realign the findings specifically to the papers titled, for the readers' comfort and benefits.  
+
+<|ref|>text<|/ref|><|det|>[[118, 616, 878, 659]]<|/det|>
+In general, the body of work completed was commendable. At the same time, the written language may not be palatable to a wider audience and readers. It would be advisable to make it simple where possible since there are several paragraphs confounded with each other.  
+
+<|ref|>text<|/ref|><|det|>[[118, 671, 876, 714]]<|/det|>
+One of the classic examples in the paper titled "Systematic Review and Meta- Analysis for a Global Patient co- Owned Cloud (GPOC)" uses the term "use- ubiquity globally" in the abstract. This term is misleading based on the conclusion.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[118, 131, 786, 147]]<|/det|>
+Please see the below answers and actions to all the 30 points of the major revision.
+
+<|ref|>text<|/ref|><|det|>[[147, 150, 866, 180]]<|/det|>
+(1) The manuscript **GPOC SYS-META Clean** does not contain comments, it is a revised clean manuscript.
+
+<|ref|>text<|/ref|><|det|>[[147, 183, 824, 212]]<|/det|>
+(2) The manuscript **GPOC SYS-META Comms** - you find the actions labelled with comments #1-30.
+
+<|ref|>text<|/ref|><|det|>[[147, 215, 880, 245]]<|/det|>
+(3) The manuscript **GPOC SYS-META Highlighted** - you find all the changes underlined and highlighted.
+
+<|ref|>text<|/ref|><|det|>[[147, 247, 850, 278]]<|/det|>
+(4) For convenience, the manuscript **GPOC SYS-META Trace** (Word-tracked) is fully traceable from the previous version that you reviewed.
+
+<|ref|>table<|/ref|><|det|>[[108, 304, 895, 875]]<|/det|>
+
+<table><tr><td>Comments</td><td>#</td><td>Part</td><td>Point by Point Answers & Actions</td></tr><tr><td>Editor comments found in the<br>separate cover letter.</td><td>1-8</td><td>All</td><td>See author cover letter.</td></tr><tr><td colspan="4">REVIEWER COMMENTS</td></tr><tr><td colspan="4">Reviewer #1 (Remarks to the<br>Author)</td></tr><tr><td>The article systematically<br>reviews sharing of electronic<br>health records in a cloud. The<br>text is original, follows<br>PRISMA protocol, and presents<br>exciting insights to the<br>scientific community. The<br>authors considered the most<br>significant articles on the<br>subject. The protocol is<br>adequately done and gives a<br>discussion pointing out the<br>main challenges and issues.</td><td>9</td><td>All</td><td>We thank the reviewer for constructive and helpful comments that has ameliorated the manuscript.</td></tr><tr><td>Following, I list some<br>suggestions to improve the<br>article: GPOC is not a<br>universally accepts term. It is a term that the authors have used<br>in previous publications.<br>Although I do not see a<br>problem using this term in the<br>article, the authors could better<br>define what it means and why it could not use just PHR to<br>represent this idea. Another<br>doubt is what is a co-owned<br>cloud. Is it a public cloud? Is it<br>co-owned by the patient and its<br>health providers?</td><td>10</td><td>Abst.<br>&<br>Intro.</td><td>It is important to define the unicity of the GPOC concept. This has now been implemented into the abstract and introduction. Hence, now there is a better definition of GPOC, what the co-ownership is, and why it is trisected for legal reasons. It has also been highlighted why a just PHR cannot be used - the<br>unicity of the GPOC concept of cloud, ownership, AI<br>integration, blockchain, foundation, globality, sharing,<br>independence, interaction, legal foundation status and<br>substrate for development of ML and its global dissemination has been elaborated.</td></tr><tr><td>- Pg. 3, line 45: The authors<br>discuss centralized approaches.<br>It would be interesting to<br>mention fog-based and peer-to-peer/hierarchical methods for<br>completeness.</td><td>11</td><td>Disc.</td><td>These have now been implemented as valuable new angles.<br>Now the fog-based and peer-to-peer/hierarchical methods are mentioned.</td></tr></table>
+
+<--- Page Split --->
+<|ref|>table<|/ref|><|det|>[[108, 80, 895, 900]]<|/det|>
+
+<table><tr><td>- Pg. 3, line 63: How to deal with the language barrier among cross-border travel? I would appreciate a discussion on this subject.</td><td>12</td><td>Disc.</td><td>This is now included in the discussion among the challenges and importance of AI integration into GPOC. We exemplify how inbuilt AI in the GPOC can solve language barriers.</td></tr><tr><td>- When discussing security and privacy, one of the focuses of the article, you could analyze the implications of HIPAA and GDPR. There are many aspects that these kinds of legislation affect privacy regarding health records and their relation to cross-border travel.</td><td>13</td><td>Disc.</td><td>Thank you, we have now added a comment about and discuss this. After this comment we refer to the GPOC Ethics article 4 in the reference list. This article focuses entirely on ethics, policy, legal, regulation, etc. This is a self-contained article within the GPOC series. GDPR &amp;amp; HIPAA now mentioned (without infringing into the Ethics' article), but now highlighted. Indeed, we cross reference to the article that focuses in this area.</td></tr><tr><td>- Security is mainly discussed in terms of encryption and decryption of data. A more comprehensive analysis is needed, including the three main aspects of confidentiality, integrity, and availability.</td><td>14</td><td>Disc.</td><td>Now included and analysed articles of the systematic review elaborating these main aspects of confidentiality (19), integrity (11), and availability (12). Within parenthesis the number of articles bringing this up.</td></tr><tr><td>- The method is well-defined and based on PRISMA. I did not find the range period for the search, the bases used, and the specific terms (on page 5, line 94, you only mention some keywords).</td><td>15</td><td>Meth. &amp;amp; Supp.</td><td>The PRISMA form and other search documents have now all been included as supplements and referred to as supplementary files S1 and S3. In S1 the full search strategy is visible, with search strings etc. In S3 the PRISMA checklist is presented.</td></tr><tr><td>- Interoperability is an important issue that is not discussed at all. What is the rationale, and why was it not considered in the article? A global trend of adopting HL7 FHIR and other standards could influence the field.</td><td>16</td><td>Disc.</td><td>Now interoperability has been implemented into the discussion. Now also HL7 FHIR is mentioned. Also, as cross referencing here has been made to the technical EcoTech Article.2</td></tr><tr><td>- Security-based parameters are based on times in ms: time for encryption, time for decryption, and the ratio of means. These times depend on the infrastructure: CPU, memory, network bandwidth, etc. Just showing the number of effective time adds little value since it depends on the infrastructure used.</td><td>17</td><td>Disc.</td><td>Now this point has been added to the big picture of the results, as a decisive factor to consider and weigh in. The infrastructure for GPOC is further elaborated in the technical EcoTech Article.2 It is cross referenced to this now too.</td></tr><tr><td>- I missed a big picture of the main issues in the discussion section. The discussion is based on the more traditional "Author X studied that; Author Y discusses." You could start with a big picture presenting the main topics and then more implicitly bringing the related studies.</td><td>18</td><td>Disc.</td><td>Now the big picture has been added in an initial overview, but also in the discussion. Al the places with the traditionally mention "author X studied Y", etc, have been altered and the result is a much better flow. Now, with an accompanying big picture before and after, it is more readable. As advised the big picture is presented mainly first. Also, the traditional author mentioning has been partly altered where possible now, so that the sentences are more neutral as flow better for the reader.</td></tr><tr><td>- You cite types of encryptions on lines 281-282 (page 14). A</td><td>19</td><td>Disc.</td><td>A discussion has now been added on encryption. Also, referring further the technical EcoTech Article.2</td></tr></table>
+
+<--- Page Split --->
+<|ref|>table<|/ref|><|det|>[[108, 80, 895, 900]]<|/det|>
+
+<table><tr><td>discussion could be added<br>because it could influence many<br>essential aspects of the study.</td><td></td><td></td><td></td></tr><tr><td>- Blockchain is cited throughout the article, and I consider it a<br>trend. You could discuss this<br>technology further and its<br>impact with more emphasis.<br>Other crucial subjects still need<br>to be explored, including<br>Federated Learning, Fog<br>Computing, and the Internet of<br>Things (how to combine data<br>from wearables from patient<br>EHR, for instance).</td><td>20</td><td>Disc.</td><td>Now a discussion on blockchain has been added. Federated Learning, Fog Computing, and the Internet of Things are mentioned, but also a refence for further reading of the technical EcoTech Article.2</td></tr><tr><td>- On page 19 you discuss<br>machine learning and the use of<br>data for decision, preserving<br>privacy, and not disclosing<br>patients' details. Anonymization and obfuscation, among other<br>methods, could be commented<br>on here. Although machine<br>learning is not a focus of the<br>article, it could influence GPOC in the services provided and<br>organization/storage of data.</td><td>21</td><td>Disc.</td><td>Now anonymization and obfuscation and also fully<br>homomorphic encryption is mentioned. There is now also a refence for further reading, where this is elaborated, in the technical EcoTech Article.2</td></tr><tr><td>- Line 386, Pg. 20, you cite<br>"four main areas," but it seems<br>to be five. It would be nice to<br>see some future directions.</td><td>22</td><td>Disc.</td><td>It has now been changed to eight (8) areas. Moreover, the initial keywords of each numbered paragraph have been italicised for increased clarity and readability. Also, the<br>number of directions has been expanded and are now also synched with the six problems statements mentioned before and also referred to in the GPOC series. Again, without any overlap.</td></tr><tr><td>Reviewer #2 (Remarks to the<br>Author)</td><td></td><td></td><td></td></tr><tr><td>I appreciate and congratulate all<br>authors for adding knowledge<br>to the Global Patient co-Owned Cloud (GPOC) a Cloud-based<br>infrastructure that shares<br>information in the Personal<br>Health Records (PHRs) space.<br>This topic is timely and requires significant effort to establish<br>particularly the concern on data security and privacy. I believe<br>the meta-analysis and results<br>presented cover the area to<br>some extent however the term<br>"use-ubiquity globally" that<br>appears in the abstract does not<br>take any advantages and/or<br>support to the claims (refer to<br>the conclusion).</td><td>23</td><td>Abst.</td><td>Many thanks. And yes, the term "use-ubiquity globally" that appears in the abstract, has been deleted and the whole<br>abstract has been completely rewritten. It is now within the<br>required limits of &lt;150 words. Decreased from 244 words.<br>The abstract now starts to explain GPOC, present the series to get a context for the reader, and then focuses on the article and what it shows.</td></tr><tr><td>Some areas to improve based<br>on the Electronic Health<br>Records (EHRs) use are: 1)<br>Validity and Bias: This is very</td><td>24</td><td>Res.</td><td>Now under the headline "Validity and Bias" it has now been<br>explained how the risk minimisation was managed.</td></tr></table>
+
+<--- Page Split --->
+<|ref|>table<|/ref|><|det|>[[108, 80, 895, 900]]<|/det|>
+
+<table><tr><td>critical to explain how the risk minimisation was managed.</td><td></td><td></td><td></td></tr><tr><td>(2) Perhaps, the language tone could be lower to support the general audience and/or readers (since this is Open Access Forum)</td><td>25</td><td>All</td><td>As the editor also requested (and referred to reviewer 2) above in comment #1, very sentence of the whole manuscript has been put under revision and all paragraphs and nearly all sentences have been altered or polished. Hence, the language is now more suitable to a general audience and the sentences are shorter.</td></tr><tr><td>(3) In addition, Blockchain protocol-embedded PHRs might be another approach that did not dominate the claim hence expanding such idea and thinking would be an added advantage to the paper.</td><td>26</td><td>Disc.</td><td>Now the approach with Blockchain protocol-embedded PHRs have been mentioned. Reference added for further reading of the technical EcoTech Article.²</td></tr><tr><td>(4) The conclusion section: I believe there are standards when sharing EHRs. Such standards are depending on the Geographical locations, the Country's data protection, and/or breaches of legislation and their maturity (this is a very challenging area indeed).</td><td>27</td><td>Disc. & Con.</td><td>Now a hint about this is added. This is the full subject and ranger of the Ethics Article⁴. This article fully delves into geographical variations of ethics, legislation, regulations and policies. It is an additional combinatory literature review and interview series on this subject. It could not have been included into the present systematic review and meta-analysis and is self-contained and stands on its own legs. But a cross referencing has been added with the mentioned hint that this is important, and worth an independent article. See also the last point 8 of the discussion summary.</td></tr><tr><td colspan="4">ADDITIONAL COMMENTS</td></tr><tr><td>The manuscripts share some form of information from the literature search. While they are interconnected, scientifically it is appropriate and valid to share the information however, the merit of the research contributions is somewhat diluted. Hence, it would be advisable to realign the findings specifically to the papers titled, for the readers' comfort and benefits.</td><td>28</td><td>Abst. & Intro</td><td>Absolutely. The overlaps have now been minimised. Hopefully eradicated. For instance, it is requested in the revision to delve into a technical or regulatory area, and these have now been added, but only very brief and with a reference to the article which has this area as its main subject. It is hence important that the required technical additions to the systematic review do not infringe or dilute other manuscripts. Note: all mentioning of subjects belonging to another manuscript has been deleted. Similarly other subjects belonging to other self-contained parts of the series have been deleted and referred to only once for information to the reader (ref 1,2,3,4). Hence, a total realignment has been carried out.</td></tr><tr><td>In general, the body of work completed was commendable. At the same time, the written language may not be palatable to a wider audience and readers. It would be advisable to make it simple where possible since there are several paragraphs confounded with each other.</td><td>29</td><td>All</td><td>The whole text has now been rewritten. The language has been improved and the text flows better. It is adapted to the general audience. The sentences are less complex and shorter. Easier terminology has been used, whenever possible. When comparing the previous version and the present article manuscript all paragraphs and a majority of sentences have been rewritten to ease the flow for the general audience, which is very important. The paragraphing has been changed to eradicate any confusion. Note that now both the abstract (in two short sentences) and the introduction's first paragraph briefly mention the GPOC concept and the context of the GPOC series. Hence, the reader will immediately know what to expect from the articles regarding scope and contents. At the onset the reader will be aware that the five GPOC series entities will display distinct and non-overlapping foci. Hence, this sets the tone and contributes to the realignment of the articles. Several paragraphs hence became superfluous and have been deleted.</td></tr></table>
+
+<--- Page Split --->
+<|ref|>table<|/ref|><|det|>[[108, 82, 890, 210]]<|/det|>
+
+<table><tr><td>One of the classic examples in the paper titled "Systematic Review and Meta-Analysis for a Global Patient co-Owned Cloud (GPOC)" uses the term "use-ubiquity globally" in the abstract. This term is misleading based on the conclusion.</td><td>30</td><td>Abst.</td><td>This term has been deleted. The abstract has been completely rewritten. It has also been shortened from 244 to the required 150 words maximum. All intricate sentences with complex syntax and prosody have been altered. The message is clearer. Sentences shorter. Language simpler.</td></tr></table>
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[118, 85, 293, 98]]<|/det|>
+Reviewers' Comments:  
+
+<|ref|>text<|/ref|><|det|>[[118, 113, 222, 125]]<|/det|>
+Reviewer #1:  
+
+<|ref|>text<|/ref|><|det|>[[118, 127, 298, 140]]<|/det|>
+Remarks to the Author:  
+
+<|ref|>text<|/ref|><|det|>[[118, 141, 853, 169]]<|/det|>
+The authors have conducted all the modifications asked for in the previous review. The article is now ready to be accepted.
+
+<--- Page Split --->

peer_reviews/126886759851d7ab006726a2a5b4c6ce9d24ee1c0468e8a9a6cfd9c1f1293206/supplementary_0_Peer Review File/images_list.json

@@ -0,0 +1 @@
+[]

peer_reviews/126886759851d7ab006726a2a5b4c6ce9d24ee1c0468e8a9a6cfd9c1f1293206/supplementary_0_Peer Review File/supplementary_0_Peer Review File.mmd

@@ -0,0 +1,188 @@
+
+# nature portfolio  
+
+Peer Review File  
+
+Supertorical light pulses as electromagnetic skyrmions propagating in free space  
+
+![PLACEHOLDER_0_0]
+  
+
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+
+## REVIEWER COMMENTS  
+
+Reviewer #1 (Remarks to the Author):  
+
+The topology of electromagnetic beams and pulses is at the focus of growing research efforts currently, in particular with respect to the possibility of studying and exploiting topological phenomena typically observed in condensed matter (e.g. skyrmions). Whereas a number of works have studied such phenomena in electromagnetic beams, not much is known about the topological properties of pulses. Thus, the presented results are important and will be of considerable interest to the community. Overall, the paper is well written, and I recommend it for publication. I also suggest the following optional improvement to the manuscript:  
+
+1. The authors may want to discuss in more detail the importance of their findings, in particular in the context of the recent stream of works (e.g. refs. [19,20,24]) on "skyrmion-like" electromagnetic field configurations. 
+2. The authors may want to qualitatively discuss the dependence of the topological structure on the q1 & q2 parameters. 
+3. The authors may want to suggest methods to generate the "supertoroidal pulses". Can they be generated in principle and what are the practical challenges? 
+4. The singularity structures presented in the Fig. 1 are not well explained. For instance, in the inset to Fig. 1a, what is annotated as line singularity appears to be a saddle point singularity.  
+
+Finally, there is a typo in the beginning of the Discussion section (line 4): "maagnetic".  
+
+Reviewer #2 (Remarks to the Author):  
+
+The authors introduce a complete theoretical work of a new family of structured light pulses, which possess many attractive topological properties, including the general toroidal topology, skyrmions, fractal patterns. Both the toroidal electrodynamics and optical skyrmions are increasingly hot topics in recent years, and I am pleased to see the presented modeling in this manuscript manage to find an intersection of these two subjects. Also, the pulse family refers to the structuring of few- cycle ultrafast pulses, thus it shows importance to open new research direction of ultrafast nonlinear optics. To me it is also the first demonstration of optical quasiparticle, i.e. the skyrmions in propagating freespace light pulses, bring new topological states of light. Therefore, this article has enough novelty that deserved to be published, after some revisions as suggested below.  
+
+(1) In my opinion, the work delivers two main novelties, the toroidal and the skyrmion. However, the backgrounds for both subjects are not included in the Introduction section. What is the importance of exploring new toroidal topology? What is the motivation of optical skyrmion? Besides, the authors have a paragraph introducing THz wave, which is confusing to me, since I don't get how THz motivate this work. Anyway, the article is well written except the Introduction section. Please reconsider the logic to give a more convincing introduction.  
+
+(2) There is a misleading citation in the last sentence of first paragraph, that reference [25] has nothing to do with spin-orbit optical forces, while it is a work on nucleon interaction. I suggest this citation being replaced or removed. I understand skyrmion is a very hot topic, but please be careful that many works are not in optics, and what discussed here is optical skyrmion.  
+
+(3) The "Results" part is very good, except a few expressions in the last subsection. For "fine-scale", how fine is the scale? Nanoscale or microscale? The author indeed explained this that it is sub-q1 scale, and q1 is effective wavelength, but one should avoid any ambiguous adjective in academic papers. Moreover, I found the sub-"wavelength' feature of skyrmion discussed here is extremely similar to a prior important work "Deep-subwavelength features of photonic skyrmions in a confined electromagnetic field with orbital angular momentum. Nat. Phys. 15, 650-654 (2019)." The similarity of two works should be discussed.  
+
+(4) The supplemental material includes a very clear and detailed derivation of supertoroidal pulses, but no basic theoretical framework on skyrmion. The Methods include a part introducing the skyrmion background,
+
+<--- Page Split --->
+
+but, as a reader, I am looking for a clearer theoretical background showing what is the ideal state of topological skyrmions. It helps if some simulation result can be provided so that one can compare the skyrmions in supertoroidal pulses with the ideal models. I suggest authors add some results in the supplemental material.  
+
+Reviewer #3 (Remarks to the Author):  
+
+The manuscript reports a family of few- cycle toroidal light pulses. The groundwork to this description was set in 1989, when Zielkowsky derived exact solution to Maxwell's equations in free space that combine properties of electromagnetic bullets and transient beams, including the defining equation of toroidal beams. Several properties of the simplest toroidal beams with azimuthal electric or magnetic fields were described soon after in 1996 by Hellwarth and Nouchi. The current manuscript extends this description to include higher order modes of toroidal light pulses, and describes phenomenologically several of their properties, including the spatial distribution of their vector fields, some considerations of the skyrmion character and energy backflow.  
+
+This is in principle an interesting development, and, without following the mathematical derivation in detail, I have no doubt that the work is correct.  
+
+What the manuscript lacks, in my opinion, is an explanation of why the described properties are important, and a motivation that goes beyond superficial hints at potential applications in information transfer or microscopy.  
+
+Where are the higher order modes superior?  
+
+What is the role of the defining parameter alpha, is there a fundamental difference between integer and fractal values of alpha, how does it impact on the skyrmion number or the energy backflow? What are general considerations beyond describing features of the specific examples? Are there experimental considerations for generating the more sophisticated modes?  
+
+I note that the "growing attention" that toroidal light pulses are apparently attracting is demonstrated with 10 papers sharing the final author with the current manuscript - a more balanced view would strengthen this point.  
+
+Finally, I suggest to replace the term "supertoroidal" with the more neutral "generalised toroidal light pulses".
+
+<--- Page Split --->
+
+## Reviewer #1:  
+
+"The topology of electromagnetic beams and pulses is at the focus of growing research efforts currently, in particular with respect to the possibility of studying and exploiting topological phenomena typically observed in condensed matter (e.g. skyrmions). Whereas a number of works have studied such phenomena in electromagnetic beams, not much is known about the topological properties of pulses. Thus, the presented results are important and will be of considerable interest to the community. Overall, the paper is well written, and I recommend it for publication. I also suggest the following optional improvement to the manuscript:"  
+
+We thank the reviewer for their positive comments and recommendation for publication.  
+
+"1. The authors may want to discuss in more detail the importance of their findings, in particular in the context of the recent stream of works (e.g. refs. [19,20,24]) on "skyrmion-like" electromagnetic field configurations."  
+
+We now include a comprehensive discussion of recent advances in the field of electromagnetic skyrmions in the Introduction, see page 1, left column, first paragraph.  
+
+"2. The authors may want to qualitatively discuss the dependence of the topological structure on the \(q1\) & \(q2\) parameters."  
+
+In response to the reviewer's comment, we discuss the dependence of the topological structure on the \(q_{1}\) and \(q_{2}\) parameters in page 7, right column, end of first paragraph. We have also added two videos in the supplementary material to illustrate this dependence (see Supplementary Videos 3 & 6).  
+
+"3. The authors may want to suggest methods to generate the "supertoroidal pulses". Can they be generated in principle and what are the practical challenges?"  
+
+The main challenges for the generation of supertoroidal pulses involve its toroidal topology, broad bandwidth (single- cycle duration), and complex spatially- dependent spectral structure (see Supplementary Information E). We argue that supertoroidal pulses can be generated similarly to the generation of fundamental toroidal pulses [30,31], i.e. by conversion of ultrashort linearly polarized pulses in a two- stage process. This process shall involve the linear- to- radial polarization conversion of an ultrashort laser pulse, followed by the spatio- spectral modification of the pulse in a multi- layered gradient metasurface. We anticipate that the requirement for the single- cycle temporal profile will be possible to be met if we use attosecond laser pulses as input. Alternatively, in the THz range, single- cycle pulses can be routinely generated by optical rectification of femtosecond optical pulses.  
+
+We discuss the generation of supertoroidal pulses in page 7, right column, \(2^{\mathrm{nd}}\) paragraph of the Discussion section.
+
+<--- Page Split --->
+
+“4. The singularity structures presented in the Fig. 1 are not well explained. For instance, in the inset to Fig. 1a, what is annotated as line singularity appears to be a saddle point singularity.”  
+
+We thank the reviewer for pointing out this error. In the revised manuscript, we have corrected and expanded the description of singularities in Fig. 1: grey lines represent vortex- type singularities, solid circles indicate saddle- type singularities, and the colored vectors represent skyrmionic structures in selected transverse planes (see page 2, last paragraph before section “Electric field singularity”).  
+
+“Finally, there is a typo in the beginning of the Discussion section (line 4): "maagnetic".”  
+
+We have corrected this typo.  
+
+## Reviewer #2:  
+
+“The authors introduce a complete theoretical work of a new family of structured light pulses, which possess many attractive topological properties, including the general toroidal topology, skyrmions, fractal patterns. Both the toroidal electrodynamics and optical skyrmions are increasingly hot topics in recent years, and I am pleased to see the presented modeling in this manuscript manage to find an intersection of these two subjects. Also, the pulse family refers to the structuring of few- cycle ultrafast pulses, thus it shows importance to open new research direction of ultrafast nonlinear optics. To me it is also the first demonstration of optical quasiparticle, i.e. the skyrmions in propagating freespace light pulses, bring new topological states of light. Therefore, this article has enough novelty that deserved to be published, after some revisions as suggested below.”  
+
+We thank the reviewer for their supportive comments.  
+
+“(1) In my opinion, the work delivers two main novelties, the toroidal and the skyrmion. However, the backgrounds for both subjects are not included in the Introduction section. What is the importance of exploring new toroidal topology? What is the motivation of optical skyrmion? Besides, the authors have a paragraph introducing THz wave, which is confusing to me, since I don’t get how THz motivate this work. Anyway, the article is well written except the Introduction section. Please reconsider the logic to give a more convincing introduction.”  
+
+We thank the reviewer for their suggestion. We have expanded our Introduction section to include a discussion of earlier work on the “toroidal” and “skyrmion” aspects of our work and have removed the paragraph on THz waves (see page 1, paragraphs 1 & 2).  
+
+“(2) There is a misleading citation in the last sentence of first paragraph, that reference [25] has nothing to do with spin-orbit optical forces, while it is a work on nucleon interaction. I suggest this citation being replaced or removed. I understand skyrmion is a very hot topic, but please be careful that many works are not in optics, and what discussed here is optical skyrmion.”  
+
+We thank the reviewer for pointing out this error, which has been corrected in the revised manuscript.
+
+<--- Page Split --->
+
+"(3) The "Results" part is very good, except a few expressions in the last subsection. For "fine-scale", how fine is the scale? Nanoscale or microscale? The author indeed explained this that it is sub- q1 scale, and q1 is effective wavelength, but one should avoid any ambiguous adjective in academic papers. Moreover, I found the sub-'wavelength' feature of skyrmion discussed here is extremely similar to a prior important work "Deep-subwavelength features of photonic skyrmions in a confined electromagnetic field with orbital angular momentum. Nat. Phys. 15, 650-654 (2019)." The similarity of two works should be discussed."  
+
+In response to the reviewer's comment, we have replaced the phrase "fine- scale" with "subwavelength" and clearly state that the "wavelength" here refers to \(\mathrm{q}_1\) , the effective wavelength (cycle- length) of the pulse.  
+
+There is a dramatic difference between the subwavelength features presented in our work and that of Du et al., as our results involve free- space propagating waves rather than evanescent plasmonic fields. This is discussed in the Section "Subwavelength features of skyrmionic structures".  
+
+"(4) The supplemental material includes a very clear and detailed derivation of supertoroidal pulses, but no basic theoretical framework on skyrmion. The Methods include a part introducing the skyrmion background, but, as a reader, I am looking for a clearer theoretical background showing what is the ideal state of topological skyrmions. It helps if some simulation result can be provided so that one can compare the skyrmions in supertoroidal pulses with the ideal models. I suggest authors add some results in the supplemental material."  
+
+We thank the reviewer for their suggestion. We have added a discussion of the theoretical framework of skyrmions in Supplementary Information F.  
+
+## Reviewer #3:  
+
+"The manuscript reports a family of few- cycle toroidal light pulses. The groundwork to this description was set in 1989, when Zielkowsky derived exact solution to Maxwell's equations in free space that combine properties of electromagnetic bullets and transient beams, including the defining equation of toroidal beams. Several properties of the simplest toroidal beams with azimuthal electric or magnetic fields were described soon after in 1996 by Hellwarth and Nouchi. The current manuscript extends this description to include higher order modes of toroidal light pulses, and describes phenomenologically several of their properties, including the spatial distribution of their vector fields, some considerations of the skyrmion character and energy backflow. This is in principle an interesting development, and, without following the mathematical derivation in detail, I have no doubt that the work is correct. What the manuscript lacks, in my opinion, is an explanation of why the described properties are important, and a motivation that goes beyond superficial hints at potential applications in information transfer or microscopy."  
+
+We thank the reviewer for their positive comments.  
+
+"Where are the higher order modes superior?"
+
+<--- Page Split --->
+
+Supertoroidal pulses exhibit a number of advantages with respect to the fundamental Flying Doughnuts. Firstly, higher order toroidal pulses exhibit a range of different skyrmionic field configurations, which will be of interest for probing the topology of excitations in matter. Secondly, information can be encoded in the increasingly complex topological structure of the propagating pulses, which could be of interest for optical communications. Finally, the subwavelength features of the singular structures of the pulses may lead to new approaches to superresolution imaging and nanoscale metrology.  
+
+We include now this discussion in the final paragraph of the Discussion section, in page 7, right column.  
+
+"What is the role of the defining parameter alpha, is there a fundamental difference between integer and fractal values of alpha, how does it impact on the skyrmion number or the energy backflow?"  
+
+The parameter \(\alpha\) prescribes the degree of energy confinement in the pulse and is directly related to the finite energy requirement for the description of physical pulses. In particular, \(\alpha < 1\) results in pulses of infinite energy, such as planar waves and cylindrical waves, while \(\alpha \geq 1\) leads to finite- energy pulses. The parameter \(\alpha\) is also related to the topology of the pulse. Indeed, given the few- cycle nature of the supertoroidal pulses, the increasing localization of energy for high values of \(\alpha\) results in increasingly complex topological structures. Finally, in our study we found no evidence of a qualitative difference between integer and fractional values of \(\alpha\) (see Supplementary Video 4).  
+
+We now include this discussion in page 1, first paragraph of the Results section.  
+
+"What are general considerations beyond describing features of the specific examples?"  
+
+We added general considerations, especially in (1) Light- matter interactions; (2) Information transfer; (3) Imaging and metrology. See more details same as the response to the first comment.  
+
+"Are there experimental considerations for generating the more sophisticated modes?"  
+
+Please see our response to point 3 by reviewer #1.  
+
+"I note that the "growing attention" that toroidal light pulses are apparently attracting is demonstrated with 10 papers sharing the final author with the current manuscript - a more balanced view would strengthen this point."  
+
+In the revised manuscript, we provide a more comprehensive introduction on toroidal light with balanced citations (see added Refs. [34,35,38- 42]).  
+
+"Finally, I suggest to replace the term "supertoroidal" with the more neutral "generalised toroidal light pulses"."
+
+<--- Page Split --->
+
+We respectfully disagree with the reviewer. In the electrodynamics community, the term “supertoroidal” refers to generalizations of toroidal excitations (see for example Photonics 6, 43 (2019), PRA 98, 023858 (2018)). We now motivate the use of the term “supertoroidal” in the second paragraph of the Introduction section.
+
+<--- Page Split --->
+
+Reviewers' Comments:  
+
+Reviewer #1:  
+
+Remarks to the Author:  
+
+I am satisfied with author corrections and replies related to my comments. I recommend this paper for publication in Nature Communications.  
+
+Reviewer #2:  
+
+Remarks to the Author:  
+
+The authors have made a good revision work and well addressed all the concerns. Thus I recommend its publication in Nature Communications.  
+
+Reviewer #3: None
+
+<--- Page Split --->
+
+## Response to Reviewer #1:  
+
+"I am satisfied with author corrections and replies related to my comments. I recommend this paper for publication in Nature Communications."  
+
+We thank the reviewer for their positive comments and recommendation for publication.  
+
+## Response to Reviewer #2:  
+
+"The authors have made a good revision work and well addressed all the concerns. Thus I recommend its publication in Nature Communications."  
+
+We thank the reviewer for their positive comments and recommendation for publication.
+
+<--- Page Split --->

peer_reviews/126886759851d7ab006726a2a5b4c6ce9d24ee1c0468e8a9a6cfd9c1f1293206/supplementary_0_Peer Review File/supplementary_0_Peer Review File_det.mmd

@@ -0,0 +1,258 @@
+<|ref|>title<|/ref|><|det|>[[61, 40, 506, 90]]<|/det|>
+# nature portfolio  
+
+<|ref|>text<|/ref|><|det|>[[69, 110, 362, 140]]<|/det|>
+Peer Review File  
+
+<|ref|>text<|/ref|><|det|>[[70, 155, 836, 212]]<|/det|>
+Supertorical light pulses as electromagnetic skyrmions propagating in free space  
+
+<|ref|>image<|/ref|><|det|>[[57, 732, 240, 782]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[250, 732, 912, 785]]<|/det|>
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[105, 84, 281, 98]]<|/det|>
+## REVIEWER COMMENTS  
+
+<|ref|>text<|/ref|><|det|>[[105, 113, 401, 128]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[105, 143, 936, 248]]<|/det|>
+The topology of electromagnetic beams and pulses is at the focus of growing research efforts currently, in particular with respect to the possibility of studying and exploiting topological phenomena typically observed in condensed matter (e.g. skyrmions). Whereas a number of works have studied such phenomena in electromagnetic beams, not much is known about the topological properties of pulses. Thus, the presented results are important and will be of considerable interest to the community. Overall, the paper is well written, and I recommend it for publication. I also suggest the following optional improvement to the manuscript:  
+
+<|ref|>text<|/ref|><|det|>[[102, 262, 936, 396]]<|/det|>
+1. The authors may want to discuss in more detail the importance of their findings, in particular in the context of the recent stream of works (e.g. refs. [19,20,24]) on "skyrmion-like" electromagnetic field configurations. 
+2. The authors may want to qualitatively discuss the dependence of the topological structure on the q1 & q2 parameters. 
+3. The authors may want to suggest methods to generate the "supertoroidal pulses". Can they be generated in principle and what are the practical challenges? 
+4. The singularity structures presented in the Fig. 1 are not well explained. For instance, in the inset to Fig. 1a, what is annotated as line singularity appears to be a saddle point singularity.  
+
+<|ref|>text<|/ref|><|det|>[[104, 411, 768, 426]]<|/det|>
+Finally, there is a typo in the beginning of the Discussion section (line 4): "maagnetic".  
+
+<|ref|>text<|/ref|><|det|>[[105, 471, 401, 485]]<|/det|>
+Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[104, 501, 939, 635]]<|/det|>
+The authors introduce a complete theoretical work of a new family of structured light pulses, which possess many attractive topological properties, including the general toroidal topology, skyrmions, fractal patterns. Both the toroidal electrodynamics and optical skyrmions are increasingly hot topics in recent years, and I am pleased to see the presented modeling in this manuscript manage to find an intersection of these two subjects. Also, the pulse family refers to the structuring of few- cycle ultrafast pulses, thus it shows importance to open new research direction of ultrafast nonlinear optics. To me it is also the first demonstration of optical quasiparticle, i.e. the skyrmions in propagating freespace light pulses, bring new topological states of light. Therefore, this article has enough novelty that deserved to be published, after some revisions as suggested below.  
+
+<|ref|>text<|/ref|><|det|>[[104, 636, 937, 725]]<|/det|>
+(1) In my opinion, the work delivers two main novelties, the toroidal and the skyrmion. However, the backgrounds for both subjects are not included in the Introduction section. What is the importance of exploring new toroidal topology? What is the motivation of optical skyrmion? Besides, the authors have a paragraph introducing THz wave, which is confusing to me, since I don't get how THz motivate this work. Anyway, the article is well written except the Introduction section. Please reconsider the logic to give a more convincing introduction.  
+
+<|ref|>text<|/ref|><|det|>[[104, 726, 930, 785]]<|/det|>
+(2) There is a misleading citation in the last sentence of first paragraph, that reference [25] has nothing to do with spin-orbit optical forces, while it is a work on nucleon interaction. I suggest this citation being replaced or removed. I understand skyrmion is a very hot topic, but please be careful that many works are not in optics, and what discussed here is optical skyrmion.  
+
+<|ref|>text<|/ref|><|det|>[[104, 786, 937, 875]]<|/det|>
+(3) The "Results" part is very good, except a few expressions in the last subsection. For "fine-scale", how fine is the scale? Nanoscale or microscale? The author indeed explained this that it is sub-q1 scale, and q1 is effective wavelength, but one should avoid any ambiguous adjective in academic papers. Moreover, I found the sub-"wavelength' feature of skyrmion discussed here is extremely similar to a prior important work "Deep-subwavelength features of photonic skyrmions in a confined electromagnetic field with orbital angular momentum. Nat. Phys. 15, 650-654 (2019)." The similarity of two works should be discussed.  
+
+<|ref|>text<|/ref|><|det|>[[101, 875, 930, 905]]<|/det|>
+(4) The supplemental material includes a very clear and detailed derivation of supertoroidal pulses, but no basic theoretical framework on skyrmion. The Methods include a part introducing the skyrmion background,
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[105, 83, 885, 142]]<|/det|>
+but, as a reader, I am looking for a clearer theoretical background showing what is the ideal state of topological skyrmions. It helps if some simulation result can be provided so that one can compare the skyrmions in supertoroidal pulses with the ideal models. I suggest authors add some results in the supplemental material.  
+
+<|ref|>text<|/ref|><|det|>[[106, 188, 401, 202]]<|/det|>
+Reviewer #3 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[104, 217, 936, 338]]<|/det|>
+The manuscript reports a family of few- cycle toroidal light pulses. The groundwork to this description was set in 1989, when Zielkowsky derived exact solution to Maxwell's equations in free space that combine properties of electromagnetic bullets and transient beams, including the defining equation of toroidal beams. Several properties of the simplest toroidal beams with azimuthal electric or magnetic fields were described soon after in 1996 by Hellwarth and Nouchi. The current manuscript extends this description to include higher order modes of toroidal light pulses, and describes phenomenologically several of their properties, including the spatial distribution of their vector fields, some considerations of the skyrmion character and energy backflow.  
+
+<|ref|>text<|/ref|><|det|>[[101, 338, 937, 361]]<|/det|>
+This is in principle an interesting development, and, without following the mathematical derivation in detail, I have no doubt that the work is correct.  
+
+<|ref|>text<|/ref|><|det|>[[105, 363, 928, 410]]<|/det|>
+What the manuscript lacks, in my opinion, is an explanation of why the described properties are important, and a motivation that goes beyond superficial hints at potential applications in information transfer or microscopy.  
+
+<|ref|>text<|/ref|><|det|>[[105, 425, 446, 440]]<|/det|>
+Where are the higher order modes superior?  
+
+<|ref|>text<|/ref|><|det|>[[105, 441, 900, 503]]<|/det|>
+What is the role of the defining parameter alpha, is there a fundamental difference between integer and fractal values of alpha, how does it impact on the skyrmion number or the energy backflow? What are general considerations beyond describing features of the specific examples? Are there experimental considerations for generating the more sophisticated modes?  
+
+<|ref|>text<|/ref|><|det|>[[105, 517, 939, 562]]<|/det|>
+I note that the "growing attention" that toroidal light pulses are apparently attracting is demonstrated with 10 papers sharing the final author with the current manuscript - a more balanced view would strengthen this point.  
+
+<|ref|>text<|/ref|><|det|>[[100, 576, 884, 607]]<|/det|>
+Finally, I suggest to replace the term "supertoroidal" with the more neutral "generalised toroidal light pulses".
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[144, 90, 246, 105]]<|/det|>
+## Reviewer #1:  
+
+<|ref|>text<|/ref|><|det|>[[144, 116, 854, 237]]<|/det|>
+"The topology of electromagnetic beams and pulses is at the focus of growing research efforts currently, in particular with respect to the possibility of studying and exploiting topological phenomena typically observed in condensed matter (e.g. skyrmions). Whereas a number of works have studied such phenomena in electromagnetic beams, not much is known about the topological properties of pulses. Thus, the presented results are important and will be of considerable interest to the community. Overall, the paper is well written, and I recommend it for publication. I also suggest the following optional improvement to the manuscript:"  
+
+<|ref|>text<|/ref|><|det|>[[144, 247, 781, 264]]<|/det|>
+We thank the reviewer for their positive comments and recommendation for publication.  
+
+<|ref|>text<|/ref|><|det|>[[144, 302, 854, 353]]<|/det|>
+"1. The authors may want to discuss in more detail the importance of their findings, in particular in the context of the recent stream of works (e.g. refs. [19,20,24]) on "skyrmion-like" electromagnetic field configurations."  
+
+<|ref|>text<|/ref|><|det|>[[144, 364, 830, 398]]<|/det|>
+We now include a comprehensive discussion of recent advances in the field of electromagnetic skyrmions in the Introduction, see page 1, left column, first paragraph.  
+
+<|ref|>text<|/ref|><|det|>[[144, 435, 853, 470]]<|/det|>
+"2. The authors may want to qualitatively discuss the dependence of the topological structure on the \(q1\) & \(q2\) parameters."  
+
+<|ref|>text<|/ref|><|det|>[[144, 480, 853, 548]]<|/det|>
+In response to the reviewer's comment, we discuss the dependence of the topological structure on the \(q_{1}\) and \(q_{2}\) parameters in page 7, right column, end of first paragraph. We have also added two videos in the supplementary material to illustrate this dependence (see Supplementary Videos 3 & 6).  
+
+<|ref|>text<|/ref|><|det|>[[144, 578, 853, 611]]<|/det|>
+"3. The authors may want to suggest methods to generate the "supertoroidal pulses". Can they be generated in principle and what are the practical challenges?"  
+
+<|ref|>text<|/ref|><|det|>[[144, 621, 853, 793]]<|/det|>
+The main challenges for the generation of supertoroidal pulses involve its toroidal topology, broad bandwidth (single- cycle duration), and complex spatially- dependent spectral structure (see Supplementary Information E). We argue that supertoroidal pulses can be generated similarly to the generation of fundamental toroidal pulses [30,31], i.e. by conversion of ultrashort linearly polarized pulses in a two- stage process. This process shall involve the linear- to- radial polarization conversion of an ultrashort laser pulse, followed by the spatio- spectral modification of the pulse in a multi- layered gradient metasurface. We anticipate that the requirement for the single- cycle temporal profile will be possible to be met if we use attosecond laser pulses as input. Alternatively, in the THz range, single- cycle pulses can be routinely generated by optical rectification of femtosecond optical pulses.  
+
+<|ref|>text<|/ref|><|det|>[[144, 804, 853, 837]]<|/det|>
+We discuss the generation of supertoroidal pulses in page 7, right column, \(2^{\mathrm{nd}}\) paragraph of the Discussion section.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[144, 89, 843, 124]]<|/det|>
+“4. The singularity structures presented in the Fig. 1 are not well explained. For instance, in the inset to Fig. 1a, what is annotated as line singularity appears to be a saddle point singularity.”  
+
+<|ref|>text<|/ref|><|det|>[[145, 134, 853, 220]]<|/det|>
+We thank the reviewer for pointing out this error. In the revised manuscript, we have corrected and expanded the description of singularities in Fig. 1: grey lines represent vortex- type singularities, solid circles indicate saddle- type singularities, and the colored vectors represent skyrmionic structures in selected transverse planes (see page 2, last paragraph before section “Electric field singularity”).  
+
+<|ref|>text<|/ref|><|det|>[[145, 258, 796, 275]]<|/det|>
+“Finally, there is a typo in the beginning of the Discussion section (line 4): "maagnetic".”  
+
+<|ref|>text<|/ref|><|det|>[[145, 285, 352, 301]]<|/det|>
+We have corrected this typo.  
+
+<|ref|>sub_title<|/ref|><|det|>[[145, 335, 246, 351]]<|/det|>
+## Reviewer #2:  
+
+<|ref|>text<|/ref|><|det|>[[144, 362, 853, 533]]<|/det|>
+“The authors introduce a complete theoretical work of a new family of structured light pulses, which possess many attractive topological properties, including the general toroidal topology, skyrmions, fractal patterns. Both the toroidal electrodynamics and optical skyrmions are increasingly hot topics in recent years, and I am pleased to see the presented modeling in this manuscript manage to find an intersection of these two subjects. Also, the pulse family refers to the structuring of few- cycle ultrafast pulses, thus it shows importance to open new research direction of ultrafast nonlinear optics. To me it is also the first demonstration of optical quasiparticle, i.e. the skyrmions in propagating freespace light pulses, bring new topological states of light. Therefore, this article has enough novelty that deserved to be published, after some revisions as suggested below.”  
+
+<|ref|>text<|/ref|><|det|>[[145, 545, 535, 561]]<|/det|>
+We thank the reviewer for their supportive comments.  
+
+<|ref|>text<|/ref|><|det|>[[144, 588, 853, 693]]<|/det|>
+“(1) In my opinion, the work delivers two main novelties, the toroidal and the skyrmion. However, the backgrounds for both subjects are not included in the Introduction section. What is the importance of exploring new toroidal topology? What is the motivation of optical skyrmion? Besides, the authors have a paragraph introducing THz wave, which is confusing to me, since I don’t get how THz motivate this work. Anyway, the article is well written except the Introduction section. Please reconsider the logic to give a more convincing introduction.”  
+
+<|ref|>text<|/ref|><|det|>[[145, 703, 853, 755]]<|/det|>
+We thank the reviewer for their suggestion. We have expanded our Introduction section to include a discussion of earlier work on the “toroidal” and “skyrmion” aspects of our work and have removed the paragraph on THz waves (see page 1, paragraphs 1 & 2).  
+
+<|ref|>text<|/ref|><|det|>[[145, 782, 853, 850]]<|/det|>
+“(2) There is a misleading citation in the last sentence of first paragraph, that reference [25] has nothing to do with spin-orbit optical forces, while it is a work on nucleon interaction. I suggest this citation being replaced or removed. I understand skyrmion is a very hot topic, but please be careful that many works are not in optics, and what discussed here is optical skyrmion.”  
+
+<|ref|>text<|/ref|><|det|>[[144, 860, 853, 895]]<|/det|>
+We thank the reviewer for pointing out this error, which has been corrected in the revised manuscript.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[144, 116, 854, 237]]<|/det|>
+"(3) The "Results" part is very good, except a few expressions in the last subsection. For "fine-scale", how fine is the scale? Nanoscale or microscale? The author indeed explained this that it is sub- q1 scale, and q1 is effective wavelength, but one should avoid any ambiguous adjective in academic papers. Moreover, I found the sub-'wavelength' feature of skyrmion discussed here is extremely similar to a prior important work "Deep-subwavelength features of photonic skyrmions in a confined electromagnetic field with orbital angular momentum. Nat. Phys. 15, 650-654 (2019)." The similarity of two works should be discussed."  
+
+<|ref|>text<|/ref|><|det|>[[145, 247, 853, 299]]<|/det|>
+In response to the reviewer's comment, we have replaced the phrase "fine- scale" with "subwavelength" and clearly state that the "wavelength" here refers to \(\mathrm{q}_1\) , the effective wavelength (cycle- length) of the pulse.  
+
+<|ref|>text<|/ref|><|det|>[[145, 309, 853, 360]]<|/det|>
+There is a dramatic difference between the subwavelength features presented in our work and that of Du et al., as our results involve free- space propagating waves rather than evanescent plasmonic fields. This is discussed in the Section "Subwavelength features of skyrmionic structures".  
+
+<|ref|>text<|/ref|><|det|>[[144, 387, 854, 491]]<|/det|>
+"(4) The supplemental material includes a very clear and detailed derivation of supertoroidal pulses, but no basic theoretical framework on skyrmion. The Methods include a part introducing the skyrmion background, but, as a reader, I am looking for a clearer theoretical background showing what is the ideal state of topological skyrmions. It helps if some simulation result can be provided so that one can compare the skyrmions in supertoroidal pulses with the ideal models. I suggest authors add some results in the supplemental material."  
+
+<|ref|>text<|/ref|><|det|>[[145, 501, 852, 536]]<|/det|>
+We thank the reviewer for their suggestion. We have added a discussion of the theoretical framework of skyrmions in Supplementary Information F.  
+
+<|ref|>sub_title<|/ref|><|det|>[[145, 564, 246, 580]]<|/det|>
+## Reviewer #3:  
+
+<|ref|>text<|/ref|><|det|>[[144, 589, 854, 814]]<|/det|>
+"The manuscript reports a family of few- cycle toroidal light pulses. The groundwork to this description was set in 1989, when Zielkowsky derived exact solution to Maxwell's equations in free space that combine properties of electromagnetic bullets and transient beams, including the defining equation of toroidal beams. Several properties of the simplest toroidal beams with azimuthal electric or magnetic fields were described soon after in 1996 by Hellwarth and Nouchi. The current manuscript extends this description to include higher order modes of toroidal light pulses, and describes phenomenologically several of their properties, including the spatial distribution of their vector fields, some considerations of the skyrmion character and energy backflow. This is in principle an interesting development, and, without following the mathematical derivation in detail, I have no doubt that the work is correct. What the manuscript lacks, in my opinion, is an explanation of why the described properties are important, and a motivation that goes beyond superficial hints at potential applications in information transfer or microscopy."  
+
+<|ref|>text<|/ref|><|det|>[[145, 824, 515, 840]]<|/det|>
+We thank the reviewer for their positive comments.  
+
+<|ref|>text<|/ref|><|det|>[[147, 859, 494, 875]]<|/det|>
+"Where are the higher order modes superior?"
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[144, 89, 854, 210]]<|/det|>
+Supertoroidal pulses exhibit a number of advantages with respect to the fundamental Flying Doughnuts. Firstly, higher order toroidal pulses exhibit a range of different skyrmionic field configurations, which will be of interest for probing the topology of excitations in matter. Secondly, information can be encoded in the increasingly complex topological structure of the propagating pulses, which could be of interest for optical communications. Finally, the subwavelength features of the singular structures of the pulses may lead to new approaches to superresolution imaging and nanoscale metrology.  
+
+<|ref|>text<|/ref|><|det|>[[144, 219, 839, 253]]<|/det|>
+We include now this discussion in the final paragraph of the Discussion section, in page 7, right column.  
+
+<|ref|>text<|/ref|><|det|>[[144, 291, 820, 343]]<|/det|>
+"What is the role of the defining parameter alpha, is there a fundamental difference between integer and fractal values of alpha, how does it impact on the skyrmion number or the energy backflow?"  
+
+<|ref|>text<|/ref|><|det|>[[144, 353, 854, 492]]<|/det|>
+The parameter \(\alpha\) prescribes the degree of energy confinement in the pulse and is directly related to the finite energy requirement for the description of physical pulses. In particular, \(\alpha < 1\) results in pulses of infinite energy, such as planar waves and cylindrical waves, while \(\alpha \geq 1\) leads to finite- energy pulses. The parameter \(\alpha\) is also related to the topology of the pulse. Indeed, given the few- cycle nature of the supertoroidal pulses, the increasing localization of energy for high values of \(\alpha\) results in increasingly complex topological structures. Finally, in our study we found no evidence of a qualitative difference between integer and fractional values of \(\alpha\) (see Supplementary Video 4).  
+
+<|ref|>text<|/ref|><|det|>[[144, 502, 721, 519]]<|/det|>
+We now include this discussion in page 1, first paragraph of the Results section.  
+
+<|ref|>text<|/ref|><|det|>[[145, 545, 790, 563]]<|/det|>
+"What are general considerations beyond describing features of the specific examples?"  
+
+<|ref|>text<|/ref|><|det|>[[144, 573, 842, 608]]<|/det|>
+We added general considerations, especially in (1) Light- matter interactions; (2) Information transfer; (3) Imaging and metrology. See more details same as the response to the first comment.  
+
+<|ref|>text<|/ref|><|det|>[[144, 635, 775, 652]]<|/det|>
+"Are there experimental considerations for generating the more sophisticated modes?"  
+
+<|ref|>text<|/ref|><|det|>[[144, 662, 505, 679]]<|/det|>
+Please see our response to point 3 by reviewer #1.  
+
+<|ref|>text<|/ref|><|det|>[[144, 707, 854, 758]]<|/det|>
+"I note that the "growing attention" that toroidal light pulses are apparently attracting is demonstrated with 10 papers sharing the final author with the current manuscript - a more balanced view would strengthen this point."  
+
+<|ref|>text<|/ref|><|det|>[[144, 769, 836, 802]]<|/det|>
+In the revised manuscript, we provide a more comprehensive introduction on toroidal light with balanced citations (see added Refs. [34,35,38- 42]).  
+
+<|ref|>text<|/ref|><|det|>[[144, 831, 797, 866]]<|/det|>
+"Finally, I suggest to replace the term "supertoroidal" with the more neutral "generalised toroidal light pulses"."
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[144, 90, 853, 158]]<|/det|>
+We respectfully disagree with the reviewer. In the electrodynamics community, the term “supertoroidal” refers to generalizations of toroidal excitations (see for example Photonics 6, 43 (2019), PRA 98, 023858 (2018)). We now motivate the use of the term “supertoroidal” in the second paragraph of the Introduction section.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[105, 84, 280, 98]]<|/det|>
+Reviewers' Comments:  
+
+<|ref|>text<|/ref|><|det|>[[105, 114, 210, 127]]<|/det|>
+Reviewer #1:  
+
+<|ref|>text<|/ref|><|det|>[[105, 130, 285, 142]]<|/det|>
+Remarks to the Author:  
+
+<|ref|>text<|/ref|><|det|>[[103, 144, 895, 174]]<|/det|>
+I am satisfied with author corrections and replies related to my comments. I recommend this paper for publication in Nature Communications.  
+
+<|ref|>text<|/ref|><|det|>[[105, 218, 210, 231]]<|/det|>
+Reviewer #2:  
+
+<|ref|>text<|/ref|><|det|>[[105, 234, 285, 247]]<|/det|>
+Remarks to the Author:  
+
+<|ref|>text<|/ref|><|det|>[[103, 248, 904, 277]]<|/det|>
+The authors have made a good revision work and well addressed all the concerns. Thus I recommend its publication in Nature Communications.  
+
+<|ref|>text<|/ref|><|det|>[[105, 323, 210, 352]]<|/det|>
+Reviewer #3: None
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[144, 90, 343, 106]]<|/det|>
+## Response to Reviewer #1:  
+
+<|ref|>text<|/ref|><|det|>[[144, 117, 853, 151]]<|/det|>
+"I am satisfied with author corrections and replies related to my comments. I recommend this paper for publication in Nature Communications."  
+
+<|ref|>text<|/ref|><|det|>[[144, 160, 840, 179]]<|/det|>
+We thank the reviewer for their positive comments and recommendation for publication.  
+
+<|ref|>sub_title<|/ref|><|det|>[[144, 206, 343, 223]]<|/det|>
+## Response to Reviewer #2:  
+
+<|ref|>text<|/ref|><|det|>[[144, 233, 800, 267]]<|/det|>
+"The authors have made a good revision work and well addressed all the concerns. Thus I recommend its publication in Nature Communications."  
+
+<|ref|>text<|/ref|><|det|>[[142, 276, 840, 295]]<|/det|>
+We thank the reviewer for their positive comments and recommendation for publication.
+
+<--- Page Split --->

peer_reviews/12696181428465838111cfb9bc03699d1113262f9765d95778f49019f4af4f2b/supplementary_0_Peer Review File/images_list.json

@@ -0,0 +1,25 @@
+[
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_0.jpg",
+    "caption": "Figure R1 Schematic of the experimental setup used to characterize the linewidths of individual comb lines.",
+    "footnote": [],
+    "bbox": [],
+    "page_idx": 0
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_2.jpg",
+    "caption": "Figure R2 a) Frequency noise spectrum of a single-mode lasing state, single-lobe comb state (Comb 1), and bifurcated two-lobe comb state (Comb 2). For the combs, the frequency noise is for the entire combs, similar to Fig.2 in the paper. b) Optical spectrum (blue) and recorded linewidth (red dot) for the single-mode lasing state. c) Optical spectrum (blue) and recorded linewidths (red dots) of individual comb lines, for the single-lobe comb state (Comb 1). d) Same as c) but for the bifurcated two-lobe comb state (Comb 2). In b)-d), the small spectral gap on the noise floor is due to the spectral filtering by the FBG filter.",
+    "footnote": [],
+    "bbox": [
+      [
+        171,
+        92,
+        826,
+        466
+      ]
+    ],
+    "page_idx": 7
+  }
+]

peer_reviews/12696181428465838111cfb9bc03699d1113262f9765d95778f49019f4af4f2b/supplementary_0_Peer Review File/supplementary_0_Peer Review File.mmd

@@ -0,0 +1,224 @@
+
+# nature portfolio  
+
+Peer Review File  
+
+Electrically empowered microcomb laser  
+
+![](images/Figure_unknown_0.jpg)
+  
+
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+
+## REVIEWER COMMENTS  
+
+Reviewer #1 (Remarks to the Author):  
+
+In the paper titled "Electrically empowered microcomb laser", the authors demonstrated an external cavity laser that can generate combs due to the combination of Kerr nonlinearity and EO modulation. This concept is novel, and the experimental results are solid. I think this paper can be accepted by Nature Communications after addressing the following issue:  
+
+1. I think some claims of the paper can be modified. Personally, I don't think this work addresses all the problems of "complex soliton initialization, high threshold, low power efficiency, and limited comb reconfigurability" for microcombs. It might be better to focus more on the dynamics of this unique laser configuration.  
+
+2. When calculating the wall-plug efficiency, I think the RF power applied to the cavity needs to be counted. One major problem for people using EO combs is the large RF power required. I think discussions regarding this problem should be added.  
+
+3. Can the author discuss more about the power of individual comb lines? What really matters in most of the applications is the comb line power, so a more detailed discussion could be helpful.  
+
+4. I feel that the statement "100% utilization of optical power fully contributing to comb generation" is strange. How to define which potion of power contribute to comb generation? For traditional microcomb this can be defined by the soliton state power/pump power, but I don't think it applies here.  
+
+5. What is the time domain pattern of this mode locked state? Is it more like a soliton or dark pulse? What is the pulse width?  
+
+6. Can the author add discussions about the tuning range of the FSR?  
+
+Reviewer #2 (Remarks to the Author):  
+
+This paper introduces a novel method of microcomb laser by hybridizing Kerr nonlinearity, EO modulation and the gain/lasing process in an EO modulated LN microresonator. The proposed scheme is novel and elegant in terms of the underlying concept, and exhibits significant performance advantages including high efficiency and a straightforward triggering process. While I do have some questions regarding the theoretical analysis detailed below, I believe a properly revised version could be considered for publication in Nature Communications.  
+
+1. Although the combination of Kerr nonlinearity, EO modulation and laser is intriguing, the concept of microcomb laser has previously been demonstrated using Kerr nonlinearity and lasing process [18,21]. Compared with these previous works, this comb spectra in this work show bifurcated spectral shapes
+
+<--- Page Split --->
+
+and narrower optical bandwidths. It would be great if the authors could provide more explanations and insights, from physics point of view, on the difference between these two microcomb laser schemes.  
+
+2. This method eliminates the need for pre-configuration compared with self-injection locking for microcomb stimulation. However, the proposed method require precise alignment of multiple resonances within a wideband between the two cavities, which seems quite challenging. The authors also mentioned, "this problem can be resolved by further optimization of the roundtrip length of the main laser cavity and introducing tunability". Compared with self-injection locking, which involves tuning the optical phase - an easy task in integrated photonics, this method requires tuning the cavity length or group index, which is actually highly challenging for on-chip photonics? Please comment on these problems and potential solutions, if any.  
+
+3. The authors identified three limitations in previous microcomb works, including complex initialization process, low power efficiency and limited reconfigurability. The proposed scheme does solve the first two problems quite obviously, but I'm not too sure what the authors refer to when mentioning "reconfigurability". Does this refer to the fast chirping depicted in Fig. 4, or the comb spacing tuning shown in Fig. 3? Since controlling comb spacing and spectral shape is well known technique in Kerr-based microcombs, I hope the authors could be more specific when mentioning "reconfigurability" and provide more explanations if possible.  
+
+4. The linewidth testing shown in Fig. 2e, which is obtained from a heterodyne measurement of the entire comb, is quite confusing to me. The authors claim that the tested linewidth is "the upper limit of the intrinsic linewidth of individual comb lines" (Page 4, line 222)? In my understanding, the noises of comb lines with higher signal powers would overwhelm those of low-power comb lines. As a result when normalizing the noise spectrum of the entire comb, one is probably seeing a larger noise contribution from higher-power lines, which are not necessarily better or worse than the lower-power lines. Please provide more references or proof to support this claim.  
+
+5. Following up on the above question, it would be really nice if the authors could characterize the linewidths of each comb lines. The intrinsic linewidth variation among different comb lines could assist the understanding of the microcomb dynamics. For Kerr microcombs and EO combs, they both show the increasing linewidths when moving from the center to sides (Ref. [1, 2] below). I am curious about the linewidth variation trend of the proposed microcomb under different states as shown in Fig. 2b and Fig. 5d. Specifically, for the 'clustered' comb state shown in Fig. 2b, Fig. S3b and d, would the linewidth increase from respective centers of different clusters?  
+
+[1] Lei, F., Ye, Z., Helgason, Ö. B., Fülöp, A., Girardi, M., & Torres- Company, V. (2022). Optical linewidth of soliton microcombs. Nature Communications, 13(1), 3161. [2] Skehan, J. C., Naveau, C., Schroder, J., & Andrekson, P. (2021). Widely tunable, low linewidth, and high power laser source using an electro-optic comb and injection-locked slave laser array. Optics Express, 29(11), 17077-17086.  
+
+6. The optical spectra presented in the experiment and simulation exhibit notable differences. Could the authors comment on the main reasons for the difference? Would the mismatch between the main laser
+
+<--- Page Split --->
+
+cavity and the racetrack resonator influence the comb state? Will the comb power keeping increasing with higher RF power as claimed in Page 4 line 234? As the authors emphasize that the cavities are dispersion- engineered, what is the influence of the dispersion of different cavities (the racetrack cavity and the main laser cavity) on the comb states? Comprehensive analyses are needed to get a better understanding of the proposed microcomb.  
+
+7. In Fig.2, the hybrid microcomb produces clustered microcombs, which are also shown in Fig. S3, while the spectral shapes in Fig. 5 are quite different. Could the authors comment on the possible origins of this difference. Is this due to the use of an optoelectronic feedback loop?  
+
+8. In Fig. 5, the authors show the feedback mode-locking of the comb laser. The concept is similar to the combination of EO combs and optoelectronic oscillators as demonstrated in previous works, such as Ref. [3-4] below. It is recommended to reference these works.  
+
+[3] Peng, H., Lei, P., Xie, X., & Chen, Z. (2021). Dynamics and timing-jitter of regenerative RF feedback assisted resonant electro-optic frequency comb. Optics Express, 29(26), 42435-42456. [4] Sakamoto, T., Kawanishi, T., & Izutsu, M. (2006). Optoelectronic oscillator using a LiNbO3 phase modulator for self-oscillating frequency comb generation. Optics letters, 31(6), 811-813.  
+
+9. There are several typos in describing Fig. 4:  
+
+Page 5, line 290-291: "Fig. 4f-h show the temporal variation of the 15-GHz beating signal..." While the temporal variations are presented in Fig. 4b- d.  
+
+Page 5, line 307- 309: "The frequency tuning range of 1.2 GHz at the modulation speed of 100 MHz (Fig. 4h) corresponds to a frequency tuning rate..." The fast-chirping results are given in Fig. 4d.  
+
+Page 5, line 314- 315: "As shown in Fig. 4e, the device exhibits a frequency tuning efficiency..." The efficiency is shown in Fig. 4h.
+
+<--- Page Split --->
+
+## Author's response to Manuscript NCOMMS-23-59501-T:  
+
+Dear Editors and Reviewers of Nature Communications,  
+
+We greatly appreciate your thorough editing of our manuscript and we thank the reviewers for their critical reading of our manuscript and making suggestions that have improved the manuscript. We have revised the manuscript wherever necessary. We list reviewers' comments in blue and our responses in black.  
+
+## Response to Reviewer 1's comments:  
+
+In the paper titled "Electrically empowered micro comb laser", the authors demonstrated an external cavity laser that can generate combs due to the combination of Kerr nonlinearity and EO modulation. This concept is novel, and the experimental results are solid. I think this paper can be accepted by Nature Communications after addressing the following issue.  
+
+RE: We thank the reviewer for the positive comments.  
+
+1. I think some claims of the paper can be modified. Personally, I don't think this work addresses all the problems of "complex soliton initialization, high threshold, low power efficiency, and limited comb reconfigurability" for microcombs. It might be better to focus more on the dynamics of this unique laser configuration.  
+
+RE: We thank the reviewer for the comment. We have revised the related sentences in the Abstract to make it clear. The analysis of the dynamics of our system is provided in detail in the supplementary information.  
+
+2. When calculating the wall-plug efficiency, I think the RF power applied to the cavity needs to be counted. One major problem for people using EO combs is the large RF power required. I think discussions regarding this problem should be added.  
+
+RE: We thank the reviewer for the comment. We understand that, for a commercial laser product, the overall wall-plug efficiency should include all power consumptions of the whole laser system to quantify its overall energy efficiency. However, such a metric does not provide insight into the essential performance of the semiconductor laser itself since a laser system always requires certain but different peripheral control circuits for proper operation that do not contribute to the final laser output power. This will make it difficult for readers to tell the fundamental laser performance and compare one with another. As such, the term of "wall-plug efficiency" is generally used to characterize the fundamental power efficiency of the diode laser itself, defined as the laser optical power compared with the electric power used to drive the laser diode. One typical example is that nearly all semiconductor lasers require temperature control, but the related power consumption is not included in the calculation of "wall-plug efficiency" in the literature.  
+
+We follow this convention in our paper. But to make it clear, we have added a sentence in the section of Comb Laser Performance to describe how the wall- plug efficiency is measured. Moreover, the detailed RF power applied to the device is provided in Fig.S3 of the supplementary information.  
+
+3. Can the author discuss more about the power of individual comb lines? What really matters in most of the applications is the comb line power, so a more detailed discussion could be helpful.  
+
+RE: We thank the reviewer for raising this question. The power of individual comb lines is in the range of 0.25—2mW. We have added this information in the section of Comb Laser Performance.  
+
+4. I feel that the statement "100% utilization of optical power fully contributing to comb generation" is strange. How to define which potion of power contributes to comb generation? For traditional microcomb this can be defined by the soliton state power/pump power, but I don't think it applies here.
+
+<--- Page Split --->
+
+RE: We thank the reviewer for the comment. We used this sentence to show the fact that, in our laser, all power of the optical wave stays in the form of comb output, in contrast to traditional microcomb in which only a small portion of the optical power is transferred to the comb from the pump laser. To make it clear, we have added a sentence to explain this in the Introduction section.  
+
+5. What is the time domain pattern of this mode-locked state? Is it more like a soliton or dark pulse? What is the pulse width?  
+
+RE: It is soliton-like pulses, as evident by the auto-correlation traces shown in the insets of Fig. 2b and Fig. 3b&c. The pulse width is estimated from the autocorrelation trace to be around 7 ps.  
+
+6. Can the author add discussions about the tuning range of the FSR?  
+
+RE: We thank the reviewer for the suggestion. The FSR of the comb can be tuned by about 1GHz. The details are provided in Fig.S3 of the supplementary information and are discussed in Section II.B.  
+
+## Response to Reviewer 2's comments:  
+
+This paper introduces a novel method of microcomb laser by hybridizing Kerr nonlinearity, EO modulation and the gain/lasing process in an EO modulated LN microresonator. The proposed scheme is novel and elegant in terms of the underlying concept, and exhibits significant performance advantages including high efficiency and a straightforward triggering process. While I do have some questions regarding the theoretical analysis detailed below, I believe a properly revised version could be considered for publication in Nature Communications. This paper introduces a novel method of microcomb laser by hybridizing Kerr nonlinearity, EO modulation and the gain/lasing process in an EO modulated LN microresonator. The proposed scheme is novel and elegant in terms of the underlying concept, and exhibits significant performance advantages including high efficiency and a straightforward triggering process. While I do have some questions regarding the theoretical analysis detailed below, I believe a properly revised version could be considered for publication in Nature Communications.  
+
+RE: We thank the reviewer for the positive comments.  
+
+1. Although the combination of Kerr nonlinearity, EO modulation and laser is intriguing, the concept of microcomb laser has previously been demonstrated using Kerr nonlinearity and lasing process [18,21]. Compared with these previous works, this comb spectra in this work show bifurcated spectral shapes and narrower optical bandwidths. It would be great if the authors could provide more explanations and insights, from physics point of view, on the difference between these two microcomb laser schemes.  
+
+RE: We thank the reviewer for the comment.  
+
+The fundamental mechanism of comb generation in our laser is very different from that in Ref.[18,21] where the self- emergence of the soliton- state relies crucially on the thermal- optic nonlinearity of the erbium- doped fiber amplifier (EDFA) to compensate for that in the nested microresonator. In contrast, our laser uses EO modulation to initiate the comb generation, Kerr nonlinearity to broaden the comb spectrum and phase lock comb lines, and the laser gain to sustain the comb operation. We have explained this point clearly in the sections of Introduction and Comb laser performance. To make it clearer, we have added a sentence in the section of Comb laser performance.  
+
+The narrower comb spectral bandwidth in our laser is simply due to the lower pump power available in our gain chip compared with an EDFA. We have provided a discussion about comb spectral bandwidth in the Discussion section. As to the bifurcated comb spectrum, its exact physical nature is not clear at this moment, which requires future exploration. We speculate that it could be likely related to the dispersion and mismatched group delay. Note that our laser can also produce single- lobe comb spectrum. The details are shown in Fig.S3 of the supplementary information, as well as Fig.5d of the main text.
+
+<--- Page Split --->
+
+2. This method eliminates the need for pre-configuration compared with self-injection locking for microcomb stimulation. However, the proposed method require precise alignment of multiple resonances within a wideband between the two cavities, which seems quite challenging. The authors also mentioned, "this problem can be resolved by further optimization of the roundtrip length of the main laser cavity and introducing tunability". Compared with self-injection locking, which involves tuning the optical phase - an easy task in integrated photonics, this method requires tuning the cavity length or group index, which is actually highly challenging for on-chip photonics? Please comment on these problems and potential solutions, if any.  
+
+RE: We thank the reviewer for raising this point. Controlling the cavity length can be realized by multiple approaches. Two examples are given below:  
+
+I) Heterogenous integration approach where the gain element is bonded on the top of the external laser cavity. In this approach, the cavity length is purely determined by the external laser cavity, which can be precisely controlled by the fabrication process.II) Tuning approach to add a group-delay tuning element into the external laser cavity. One example is Ref.[1] given below.[1] Y. Liu, et al, "Continuously tunable silicon optical true-time delay lines with a large delay tuning range and a low delay fluctuation," Opt. Express 32, 7848 (2024).  
+
+We have added these points in the section of Discussion.  
+
+3. The authors identified three limitations in previous microcomb works, including complex initialization process, low power efficiency and limited reconfigurability. The proposed scheme does solve the first two problems quite obviously, but I'm not too sure what the authors refer to when mentioning "reconfigurability". Does this refer to the fast chirping depicted in Fig. 4, or the comb spacing tuning shown in Fig. 3? Since controlling comb spacing and spectral shape is well known technique in Kerr-based microcombs, I hope the authors could be more specific when mentioning "reconfigurability" and provide more explanations if possible.  
+
+RE: We thank the reviewer for the comment. The reconfigurability refers both the fast chirping and comb spacing switching and tuning. For fast chirping, we have demonstrated unprecedented chirping rate up to \(2.4 \times 10^{17} \mathrm{~Hz} / \mathrm{s}\) orders of magnitude faster than other approaches. For comb spacing switching and tuning, we achieved it with high-speed EO modulation without involving complex tuning dynamics, which, again, is significantly faster than conventional Kerr microcombs. To make it clear, we have revised the related sentences in Abstract and Introduction.  
+
+4. The linewidth testing shown in Fig. 2e, which is obtained from a heterodyne measurement of the entire comb, is quite confusing to me. The authors claim that the tested linewidth is "the upper limit of the intrinsic linewidth of individual comb lines" (Page 4, line 222)? In my understanding, the noises of comb lines with higher signal powers would overwhelm those of low-power comb lines. As a result when normalizing the noise spectrum of the entire comb, one is probably seeing a larger noise contribution from higher-power lines, which are not necessarily better or worse than the lower-power lines. Please provide more references or proof to support this claim.  
+
+RE: We thank the reviewer for raising this point. The reviewer is correct and we have revised the relevant sentences in the paper. We provide more detailed linewidth characterizations in the following.  
+
+5. Following up on the above question, it would be really nice if the authors could characterize the linewidths of each comb line. The intrinsic linewidth variation among different comb lines could assist the understanding of the microcomb dynamics. For Kerr microcombs and EO combs, they both show the increasing linewidths when moving from the center to sides (Ref. [1, 2] below). I am curious about the linewidth variation trend of the proposed microcomb under different states as shown in Fig. 2b and
+
+<--- Page Split --->
+
+Fig. 5d. Specifically, for the ‘clustered’ comb state shown in Fig. 2b, Fig. S3b and d, would the linewidth increase from respective centers of different clusters?  
+
+[1] Lei, F., Ye, Z., Helgason, Ó. B., Fülöp, A., Girardi, M., & Torres- Company, V. (2022). Optical linewidth of soliton microcombs. Nature Communications, 13(1), 3161.  
+
+[2] Skehan, J. C., Naveau, C., Schroder, J., & Andrekson, P. (2021). Widely tunable, low linewidth, and high power laser source using an electro-optic comb and injection- locked slave laser array. Optics Express, 29(11), 17077- 17086.  
+
+RE: We thank the reviewer for the suggestion. We have performed detailed characterizations on the linewidths of individual comb lines. We used the setup shown in Fig. R1 below for the measurements, in which an individual comb line is separated from the rest of the comb by a tunable narrow- band fiber Bragg grating (FBG) filter and its linewidth is measured by the self- heterodyning method.  
+
+![](images/Figure_2.jpg)
+
+<center>Figure R1 Schematic of the experimental setup used to characterize the linewidths of individual comb lines. </center>  
+
+Figure R2a shows the overall frequency noise spectra for different lasing states. For the comb states, the linewidth measurement was performed on the entire combs similar to Fig. 2 in the paper. It shows an overall linewidth of about \(7\mathrm{kHz}\) , \(7\mathrm{kHz}\) , and \(1.5\mathrm{kHz}\) , respectively, for the single- mode lasing state, the single- lobe comb state, and the bifurcated two- lobe comb state. These values are slightly worse than those shown in the paper simply due to the slight degradation of the LN device and the RSOA in the past four months.  
+
+The recorded linewidths for individual comb lines are shown in Fig. R2c and d for the two comb states. Due to the limited powers of the individual comb lines, we can only perform linewidth characterizations for the central portion of the combs. Figure R2c shows that, for the single- lobe comb state, the individual comb lines exhibit linewidths in the range of \(1 - 3\mathrm{kHz}\) . Figure R2d shows that one lobe of the comb exhibits individual comb linewidths in the range of \(0.8 - 2\mathrm{kHz}\) , while the other lobe exhibits slightly larger linewidths in the range of \(5 - 8\mathrm{kHz}\) . The exact physical reason for this difference is not clear at this moment, which requires further exploration.  
+
+Overall, the linewidths of individual comb lines do not show a certain deterministic trend across the comb spectrum, distinctive to the ones shown in Ref. [1,2] pointed out by the reviewer. This is understandable since the linewidth feature shown in Ref.[1] is due to the recoil effect between Raman- induced self- frequency shift and dispersive wave generation, and that shown in Ref.[2] is due to the noise multiplication during the cascaded sideband creation in a pure EO comb. The mechanism of comb generation is fundamentally different in our comb laser which could lead to different linewidth behaviors.
+
+<--- Page Split --->
+![PLACEHOLDER_8_0]
+
+<center>Figure R2 a) Frequency noise spectrum of a single-mode lasing state, single-lobe comb state (Comb 1), and bifurcated two-lobe comb state (Comb 2). For the combs, the frequency noise is for the entire combs, similar to Fig.2 in the paper. b) Optical spectrum (blue) and recorded linewidth (red dot) for the single-mode lasing state. c) Optical spectrum (blue) and recorded linewidths (red dots) of individual comb lines, for the single-lobe comb state (Comb 1). d) Same as c) but for the bifurcated two-lobe comb state (Comb 2). In b)-d), the small spectral gap on the noise floor is due to the spectral filtering by the FBG filter. </center>  
+
+6. The optical spectra presented in the experiment and simulation exhibit notable differences. Could the authors comment on the main reasons for the difference? Would the mismatch between the main laser cavity and the racetrack resonator influence the comb state? Will the comb power keeping increasing with higher RF power as claimed in Page 4 line 234? As the authors emphasize that the cavities are dispersion-engineered, what is the influence of the dispersion of different cavities (the racetrack cavity and the main laser cavity) on the comb states? Comprehensive analyses are needed to get a better understanding of the proposed microcomb.  
+
+RE: We thank the reviewer for the detailed questions. We answer the questions separately in the following:  
+
+I) Could the authors comment on the main reasons for the difference?  
+
+In fact, we are able to generate a comb from Laser \(\beta\) similar to the simulation, as we showed in Fig.S2 of the supplementary information. The mechanism underlying the bifurcated comb produced by Laser \(\alpha\) is currently still under investigation, which may result from complex physics not included in the simplified model. We have added one sentence in Section III.B of the supplementary information to make this point clear.  
+
+II) Would the mismatch between the main laser cavity and the racetrack resonator influence the comb state?
+
+<--- Page Split --->
+
+Yes, the mismatch has strong impact on the stability and shape of the comb state. Chaotic comb may occur in a mismatched laser. We are currently still investigating the dynamics of the laser at different cavity length offset.  
+
+III) Will the comb power keeping increasing with higher RF power as claimed in Page 4 line 234?  
+
+The detailed relationship between the produced comb laser power and the RF driving power is shown in Fig. S3 a,c,f,h of the supplementary information for the two lasers. We observed comb power saturation in Laser \(\beta\) . In Laser \(\alpha\) , however, we haven't observed saturation within the applicable RF power range.  
+
+IV) As the authors emphasize that the cavities are dispersion-engineered, what is the influence of the dispersion of different cavities (the racetrack cavity and the main laser cavity) on the comb states?  
+
+Dispersion of the racetrack microresonator is crucial for the comb generation and mode locking. A slight anomalous dispersion is required, as we showed in the numerical modeling. The dispersion of the main laser cavity plays a minor role. We engineered it to be close to zero.  
+
+7. In Fig.2, the hybrid microcomb produces clustered microcombs, which are also shown in Fig. S3, while the spectral shapes in Fig. 5 are quite different. Could the authors comment on the possible origins of this difference. Is this due to the use of an optoelectronic feedback loop?  
+
+RE: We thank the reviewer for the question. We do observe different comb spectra for RF driven comb and feedback lock comb. Currently, the feedback locked comb is wider in spectrum. This could be related to that the dynamically adjusted feedbacked RF modulation assist in the generation of broader comb, but the detailed mechanism require further investigation in future works.  
+
+8. In Fig. 5, the authors show the feedback mode-locking of the comb laser. The concept is similar to the combination of EO combs and optoelectronic oscillators as demonstrated in previous works, such as Ref. [3-4] below. It is recommended to reference these works.  
+
+[3] Peng, H., Lei, P., Xie, X., & Chen, Z. (2021). Dynamics and timing-jitter of regenerative RF feedback assisted resonant electro-optic frequency comb. Optics Express, 29(26), 42435-42456.  
+
+[4] Sakamoto, T., Kawanishi, T., & Izutsu, M. (2006). Optoelectronic oscillator using a LiNbO3 phase modulator for self-oscillating frequency comb generation. Optics letters, 31(6), 811-813.  
+
+RE: We thank the reviewer for the suggestion. We have added these two references as Ref.[47, 48] in the paper.  
+
+9. There are several typos in describing Fig. 4: Page 5, line 290-291: "Fig. 4f-h show the temporal variation of the 15-GHz beating signal..." While the temporal variations are presented in Fig. 4b-d. Page 5, line 307-309: "The frequency tuning range of 1.2 GHz at the modulation speed of 100 MHz (Fig. 4h) corresponds to a frequency tuning rate..." The fast-chirping results are given in Fig. 4d. Page 5, line 314-315: "As shown in Fig. 4e, the device exhibits a frequency tuning efficiency..." The efficiency is shown in Fig. 4h.  
+
+RE: We thank the reviewer for finding these typos. We have corrected them in the paper.
+
+<--- Page Split --->
+
+## REVIEWERS' COMMENTS  
+
+Reviewer #1 (Remarks to the Author):  
+
+The authors have addressed all my comments and I suggest publication for this paper.  
+
+Reviewer #2 (Remarks to the Author):  
+
+The authors have addressed most of my comments. There is still a lot of physics here unexplored and not yet fully understood but I think it is reasonable to leave them to future studies as the results here are quite interesting and deserve to be published. There is one typo though: The unit of the phase noise in Fig. 5g should be "dBc/Hz".
+
+<--- Page Split --->

peer_reviews/12696181428465838111cfb9bc03699d1113262f9765d95778f49019f4af4f2b/supplementary_0_Peer Review File/supplementary_0_Peer Review File_det.mmd

@@ -0,0 +1,309 @@
+<|ref|>title<|/ref|><|det|>[[61, 40, 506, 90]]<|/det|>
+# nature portfolio  
+
+<|ref|>text<|/ref|><|det|>[[68, 110, 362, 140]]<|/det|>
+Peer Review File  
+
+<|ref|>text<|/ref|><|det|>[[70, 155, 630, 180]]<|/det|>
+Electrically empowered microcomb laser  
+
+<|ref|>image<|/ref|><|det|>[[57, 732, 239, 782]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[250, 732, 912, 785]]<|/det|>
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[116, 90, 291, 107]]<|/det|>
+## REVIEWER COMMENTS  
+
+<|ref|>text<|/ref|><|det|>[[116, 127, 392, 143]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 163, 856, 236]]<|/det|>
+In the paper titled "Electrically empowered microcomb laser", the authors demonstrated an external cavity laser that can generate combs due to the combination of Kerr nonlinearity and EO modulation. This concept is novel, and the experimental results are solid. I think this paper can be accepted by Nature Communications after addressing the following issue:  
+
+<|ref|>text<|/ref|><|det|>[[115, 254, 872, 327]]<|/det|>
+1. I think some claims of the paper can be modified. Personally, I don't think this work addresses all the problems of "complex soliton initialization, high threshold, low power efficiency, and limited comb reconfigurability" for microcombs. It might be better to focus more on the dynamics of this unique laser configuration.  
+
+<|ref|>text<|/ref|><|det|>[[115, 346, 833, 400]]<|/det|>
+2. When calculating the wall-plug efficiency, I think the RF power applied to the cavity needs to be counted. One major problem for people using EO combs is the large RF power required. I think discussions regarding this problem should be added.  
+
+<|ref|>text<|/ref|><|det|>[[115, 419, 880, 455]]<|/det|>
+3. Can the author discuss more about the power of individual comb lines? What really matters in most of the applications is the comb line power, so a more detailed discussion could be helpful.  
+
+<|ref|>text<|/ref|><|det|>[[115, 474, 874, 528]]<|/det|>
+4. I feel that the statement "100% utilization of optical power fully contributing to comb generation" is strange. How to define which potion of power contribute to comb generation? For traditional microcomb this can be defined by the soliton state power/pump power, but I don't think it applies here.  
+
+<|ref|>text<|/ref|><|det|>[[115, 547, 848, 582]]<|/det|>
+5. What is the time domain pattern of this mode locked state? Is it more like a soliton or dark pulse? What is the pulse width?  
+
+<|ref|>text<|/ref|><|det|>[[115, 602, 620, 619]]<|/det|>
+6. Can the author add discussions about the tuning range of the FSR?  
+
+<|ref|>text<|/ref|><|det|>[[116, 675, 392, 691]]<|/det|>
+Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 711, 875, 820]]<|/det|>
+This paper introduces a novel method of microcomb laser by hybridizing Kerr nonlinearity, EO modulation and the gain/lasing process in an EO modulated LN microresonator. The proposed scheme is novel and elegant in terms of the underlying concept, and exhibits significant performance advantages including high efficiency and a straightforward triggering process. While I do have some questions regarding the theoretical analysis detailed below, I believe a properly revised version could be considered for publication in Nature Communications.  
+
+<|ref|>text<|/ref|><|det|>[[115, 840, 864, 894]]<|/det|>
+1. Although the combination of Kerr nonlinearity, EO modulation and laser is intriguing, the concept of microcomb laser has previously been demonstrated using Kerr nonlinearity and lasing process [18,21]. Compared with these previous works, this comb spectra in this work show bifurcated spectral shapes
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 90, 870, 126]]<|/det|>
+and narrower optical bandwidths. It would be great if the authors could provide more explanations and insights, from physics point of view, on the difference between these two microcomb laser schemes.  
+
+<|ref|>text<|/ref|><|det|>[[114, 144, 881, 291]]<|/det|>
+2. This method eliminates the need for pre-configuration compared with self-injection locking for microcomb stimulation. However, the proposed method require precise alignment of multiple resonances within a wideband between the two cavities, which seems quite challenging. The authors also mentioned, "this problem can be resolved by further optimization of the roundtrip length of the main laser cavity and introducing tunability". Compared with self-injection locking, which involves tuning the optical phase - an easy task in integrated photonics, this method requires tuning the cavity length or group index, which is actually highly challenging for on-chip photonics? Please comment on these problems and potential solutions, if any.  
+
+<|ref|>text<|/ref|><|det|>[[114, 309, 872, 438]]<|/det|>
+3. The authors identified three limitations in previous microcomb works, including complex initialization process, low power efficiency and limited reconfigurability. The proposed scheme does solve the first two problems quite obviously, but I'm not too sure what the authors refer to when mentioning "reconfigurability". Does this refer to the fast chirping depicted in Fig. 4, or the comb spacing tuning shown in Fig. 3? Since controlling comb spacing and spectral shape is well known technique in Kerr-based microcombs, I hope the authors could be more specific when mentioning "reconfigurability" and provide more explanations if possible.  
+
+<|ref|>text<|/ref|><|det|>[[114, 456, 881, 584]]<|/det|>
+4. The linewidth testing shown in Fig. 2e, which is obtained from a heterodyne measurement of the entire comb, is quite confusing to me. The authors claim that the tested linewidth is "the upper limit of the intrinsic linewidth of individual comb lines" (Page 4, line 222)? In my understanding, the noises of comb lines with higher signal powers would overwhelm those of low-power comb lines. As a result when normalizing the noise spectrum of the entire comb, one is probably seeing a larger noise contribution from higher-power lines, which are not necessarily better or worse than the lower-power lines. Please provide more references or proof to support this claim.  
+
+<|ref|>text<|/ref|><|det|>[[114, 601, 878, 729]]<|/det|>
+5. Following up on the above question, it would be really nice if the authors could characterize the linewidths of each comb lines. The intrinsic linewidth variation among different comb lines could assist the understanding of the microcomb dynamics. For Kerr microcombs and EO combs, they both show the increasing linewidths when moving from the center to sides (Ref. [1, 2] below). I am curious about the linewidth variation trend of the proposed microcomb under different states as shown in Fig. 2b and Fig. 5d. Specifically, for the 'clustered' comb state shown in Fig. 2b, Fig. S3b and d, would the linewidth increase from respective centers of different clusters?  
+
+<|ref|>text<|/ref|><|det|>[[114, 747, 880, 840]]<|/det|>
+[1] Lei, F., Ye, Z., Helgason, Ö. B., Fülöp, A., Girardi, M., & Torres- Company, V. (2022). Optical linewidth of soliton microcombs. Nature Communications, 13(1), 3161. [2] Skehan, J. C., Naveau, C., Schroder, J., & Andrekson, P. (2021). Widely tunable, low linewidth, and high power laser source using an electro-optic comb and injection-locked slave laser array. Optics Express, 29(11), 17077-17086.  
+
+<|ref|>text<|/ref|><|det|>[[114, 858, 875, 894]]<|/det|>
+6. The optical spectra presented in the experiment and simulation exhibit notable differences. Could the authors comment on the main reasons for the difference? Would the mismatch between the main laser
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 90, 866, 180]]<|/det|>
+cavity and the racetrack resonator influence the comb state? Will the comb power keeping increasing with higher RF power as claimed in Page 4 line 234? As the authors emphasize that the cavities are dispersion- engineered, what is the influence of the dispersion of different cavities (the racetrack cavity and the main laser cavity) on the comb states? Comprehensive analyses are needed to get a better understanding of the proposed microcomb.  
+
+<|ref|>text<|/ref|><|det|>[[115, 200, 876, 253]]<|/det|>
+7. In Fig.2, the hybrid microcomb produces clustered microcombs, which are also shown in Fig. S3, while the spectral shapes in Fig. 5 are quite different. Could the authors comment on the possible origins of this difference. Is this due to the use of an optoelectronic feedback loop?  
+
+<|ref|>text<|/ref|><|det|>[[115, 273, 876, 326]]<|/det|>
+8. In Fig. 5, the authors show the feedback mode-locking of the comb laser. The concept is similar to the combination of EO combs and optoelectronic oscillators as demonstrated in previous works, such as Ref. [3-4] below. It is recommended to reference these works.  
+
+<|ref|>text<|/ref|><|det|>[[115, 345, 852, 417]]<|/det|>
+[3] Peng, H., Lei, P., Xie, X., & Chen, Z. (2021). Dynamics and timing-jitter of regenerative RF feedback assisted resonant electro-optic frequency comb. Optics Express, 29(26), 42435-42456. [4] Sakamoto, T., Kawanishi, T., & Izutsu, M. (2006). Optoelectronic oscillator using a LiNbO3 phase modulator for self-oscillating frequency comb generation. Optics letters, 31(6), 811-813.  
+
+<|ref|>text<|/ref|><|det|>[[115, 437, 448, 453]]<|/det|>
+9. There are several typos in describing Fig. 4:  
+
+<|ref|>text<|/ref|><|det|>[[115, 456, 860, 490]]<|/det|>
+Page 5, line 290-291: "Fig. 4f-h show the temporal variation of the 15-GHz beating signal..." While the temporal variations are presented in Fig. 4b- d.  
+
+<|ref|>text<|/ref|><|det|>[[115, 492, 867, 526]]<|/det|>
+Page 5, line 307- 309: "The frequency tuning range of 1.2 GHz at the modulation speed of 100 MHz (Fig. 4h) corresponds to a frequency tuning rate..." The fast-chirping results are given in Fig. 4d.  
+
+<|ref|>text<|/ref|><|det|>[[115, 528, 833, 562]]<|/det|>
+Page 5, line 314- 315: "As shown in Fig. 4e, the device exhibits a frequency tuning efficiency..." The efficiency is shown in Fig. 4h.
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[115, 91, 678, 112]]<|/det|>
+## Author's response to Manuscript NCOMMS-23-59501-T:  
+
+<|ref|>text<|/ref|><|det|>[[115, 128, 524, 144]]<|/det|>
+Dear Editors and Reviewers of Nature Communications,  
+
+<|ref|>text<|/ref|><|det|>[[115, 161, 882, 214]]<|/det|>
+We greatly appreciate your thorough editing of our manuscript and we thank the reviewers for their critical reading of our manuscript and making suggestions that have improved the manuscript. We have revised the manuscript wherever necessary. We list reviewers' comments in blue and our responses in black.  
+
+<|ref|>sub_title<|/ref|><|det|>[[116, 236, 424, 254]]<|/det|>
+## Response to Reviewer 1's comments:  
+
+<|ref|>text<|/ref|><|det|>[[115, 261, 882, 325]]<|/det|>
+In the paper titled "Electrically empowered micro comb laser", the authors demonstrated an external cavity laser that can generate combs due to the combination of Kerr nonlinearity and EO modulation. This concept is novel, and the experimental results are solid. I think this paper can be accepted by Nature Communications after addressing the following issue.  
+
+<|ref|>text<|/ref|><|det|>[[115, 332, 505, 349]]<|/det|>
+RE: We thank the reviewer for the positive comments.  
+
+<|ref|>text<|/ref|><|det|>[[115, 355, 882, 420]]<|/det|>
+1. I think some claims of the paper can be modified. Personally, I don't think this work addresses all the problems of "complex soliton initialization, high threshold, low power efficiency, and limited comb reconfigurability" for microcombs. It might be better to focus more on the dynamics of this unique laser configuration.  
+
+<|ref|>text<|/ref|><|det|>[[115, 426, 882, 460]]<|/det|>
+RE: We thank the reviewer for the comment. We have revised the related sentences in the Abstract to make it clear. The analysis of the dynamics of our system is provided in detail in the supplementary information.  
+
+<|ref|>text<|/ref|><|det|>[[115, 466, 882, 515]]<|/det|>
+2. When calculating the wall-plug efficiency, I think the RF power applied to the cavity needs to be counted. One major problem for people using EO combs is the large RF power required. I think discussions regarding this problem should be added.  
+
+<|ref|>text<|/ref|><|det|>[[115, 521, 882, 682]]<|/det|>
+RE: We thank the reviewer for the comment. We understand that, for a commercial laser product, the overall wall-plug efficiency should include all power consumptions of the whole laser system to quantify its overall energy efficiency. However, such a metric does not provide insight into the essential performance of the semiconductor laser itself since a laser system always requires certain but different peripheral control circuits for proper operation that do not contribute to the final laser output power. This will make it difficult for readers to tell the fundamental laser performance and compare one with another. As such, the term of "wall-plug efficiency" is generally used to characterize the fundamental power efficiency of the diode laser itself, defined as the laser optical power compared with the electric power used to drive the laser diode. One typical example is that nearly all semiconductor lasers require temperature control, but the related power consumption is not included in the calculation of "wall-plug efficiency" in the literature.  
+
+<|ref|>text<|/ref|><|det|>[[115, 689, 882, 738]]<|/det|>
+We follow this convention in our paper. But to make it clear, we have added a sentence in the section of Comb Laser Performance to describe how the wall- plug efficiency is measured. Moreover, the detailed RF power applied to the device is provided in Fig.S3 of the supplementary information.  
+
+<|ref|>text<|/ref|><|det|>[[115, 745, 882, 777]]<|/det|>
+3. Can the author discuss more about the power of individual comb lines? What really matters in most of the applications is the comb line power, so a more detailed discussion could be helpful.  
+
+<|ref|>text<|/ref|><|det|>[[115, 784, 881, 816]]<|/det|>
+RE: We thank the reviewer for raising this question. The power of individual comb lines is in the range of 0.25—2mW. We have added this information in the section of Comb Laser Performance.  
+
+<|ref|>text<|/ref|><|det|>[[115, 823, 882, 872]]<|/det|>
+4. I feel that the statement "100% utilization of optical power fully contributing to comb generation" is strange. How to define which potion of power contributes to comb generation? For traditional microcomb this can be defined by the soliton state power/pump power, but I don't think it applies here.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 90, 882, 154]]<|/det|>
+RE: We thank the reviewer for the comment. We used this sentence to show the fact that, in our laser, all power of the optical wave stays in the form of comb output, in contrast to traditional microcomb in which only a small portion of the optical power is transferred to the comb from the pump laser. To make it clear, we have added a sentence to explain this in the Introduction section.  
+
+<|ref|>text<|/ref|><|det|>[[115, 161, 881, 194]]<|/det|>
+5. What is the time domain pattern of this mode-locked state? Is it more like a soliton or dark pulse? What is the pulse width?  
+
+<|ref|>text<|/ref|><|det|>[[115, 201, 881, 233]]<|/det|>
+RE: It is soliton-like pulses, as evident by the auto-correlation traces shown in the insets of Fig. 2b and Fig. 3b&c. The pulse width is estimated from the autocorrelation trace to be around 7 ps.  
+
+<|ref|>text<|/ref|><|det|>[[115, 241, 625, 257]]<|/det|>
+6. Can the author add discussions about the tuning range of the FSR?  
+
+<|ref|>text<|/ref|><|det|>[[115, 264, 882, 297]]<|/det|>
+RE: We thank the reviewer for the suggestion. The FSR of the comb can be tuned by about 1GHz. The details are provided in Fig.S3 of the supplementary information and are discussed in Section II.B.  
+
+<|ref|>sub_title<|/ref|><|det|>[[115, 348, 424, 365]]<|/det|>
+## Response to Reviewer 2's comments:  
+
+<|ref|>text<|/ref|><|det|>[[115, 372, 882, 548]]<|/det|>
+This paper introduces a novel method of microcomb laser by hybridizing Kerr nonlinearity, EO modulation and the gain/lasing process in an EO modulated LN microresonator. The proposed scheme is novel and elegant in terms of the underlying concept, and exhibits significant performance advantages including high efficiency and a straightforward triggering process. While I do have some questions regarding the theoretical analysis detailed below, I believe a properly revised version could be considered for publication in Nature Communications. This paper introduces a novel method of microcomb laser by hybridizing Kerr nonlinearity, EO modulation and the gain/lasing process in an EO modulated LN microresonator. The proposed scheme is novel and elegant in terms of the underlying concept, and exhibits significant performance advantages including high efficiency and a straightforward triggering process. While I do have some questions regarding the theoretical analysis detailed below, I believe a properly revised version could be considered for publication in Nature Communications.  
+
+<|ref|>text<|/ref|><|det|>[[115, 556, 508, 571]]<|/det|>
+RE: We thank the reviewer for the positive comments.  
+
+<|ref|>text<|/ref|><|det|>[[117, 579, 882, 658]]<|/det|>
+1. Although the combination of Kerr nonlinearity, EO modulation and laser is intriguing, the concept of microcomb laser has previously been demonstrated using Kerr nonlinearity and lasing process [18,21]. Compared with these previous works, this comb spectra in this work show bifurcated spectral shapes and narrower optical bandwidths. It would be great if the authors could provide more explanations and insights, from physics point of view, on the difference between these two microcomb laser schemes.  
+
+<|ref|>text<|/ref|><|det|>[[115, 666, 440, 681]]<|/det|>
+RE: We thank the reviewer for the comment.  
+
+<|ref|>text<|/ref|><|det|>[[115, 690, 882, 802]]<|/det|>
+The fundamental mechanism of comb generation in our laser is very different from that in Ref.[18,21] where the self- emergence of the soliton- state relies crucially on the thermal- optic nonlinearity of the erbium- doped fiber amplifier (EDFA) to compensate for that in the nested microresonator. In contrast, our laser uses EO modulation to initiate the comb generation, Kerr nonlinearity to broaden the comb spectrum and phase lock comb lines, and the laser gain to sustain the comb operation. We have explained this point clearly in the sections of Introduction and Comb laser performance. To make it clearer, we have added a sentence in the section of Comb laser performance.  
+
+<|ref|>text<|/ref|><|det|>[[115, 810, 883, 904]]<|/det|>
+The narrower comb spectral bandwidth in our laser is simply due to the lower pump power available in our gain chip compared with an EDFA. We have provided a discussion about comb spectral bandwidth in the Discussion section. As to the bifurcated comb spectrum, its exact physical nature is not clear at this moment, which requires future exploration. We speculate that it could be likely related to the dispersion and mismatched group delay. Note that our laser can also produce single- lobe comb spectrum. The details are shown in Fig.S3 of the supplementary information, as well as Fig.5d of the main text.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 89, 882, 219]]<|/det|>
+2. This method eliminates the need for pre-configuration compared with self-injection locking for microcomb stimulation. However, the proposed method require precise alignment of multiple resonances within a wideband between the two cavities, which seems quite challenging. The authors also mentioned, "this problem can be resolved by further optimization of the roundtrip length of the main laser cavity and introducing tunability". Compared with self-injection locking, which involves tuning the optical phase - an easy task in integrated photonics, this method requires tuning the cavity length or group index, which is actually highly challenging for on-chip photonics? Please comment on these problems and potential solutions, if any.  
+
+<|ref|>text<|/ref|><|det|>[[115, 225, 881, 258]]<|/det|>
+RE: We thank the reviewer for raising this point. Controlling the cavity length can be realized by multiple approaches. Two examples are given below:  
+
+<|ref|>text<|/ref|><|det|>[[142, 264, 882, 396]]<|/det|>
+I) Heterogenous integration approach where the gain element is bonded on the top of the external laser cavity. In this approach, the cavity length is purely determined by the external laser cavity, which can be precisely controlled by the fabrication process.II) Tuning approach to add a group-delay tuning element into the external laser cavity. One example is Ref.[1] given below.[1] Y. Liu, et al, "Continuously tunable silicon optical true-time delay lines with a large delay tuning range and a low delay fluctuation," Opt. Express 32, 7848 (2024).  
+
+<|ref|>text<|/ref|><|det|>[[115, 400, 525, 416]]<|/det|>
+We have added these points in the section of Discussion.  
+
+<|ref|>text<|/ref|><|det|>[[115, 423, 882, 536]]<|/det|>
+3. The authors identified three limitations in previous microcomb works, including complex initialization process, low power efficiency and limited reconfigurability. The proposed scheme does solve the first two problems quite obviously, but I'm not too sure what the authors refer to when mentioning "reconfigurability". Does this refer to the fast chirping depicted in Fig. 4, or the comb spacing tuning shown in Fig. 3? Since controlling comb spacing and spectral shape is well known technique in Kerr-based microcombs, I hope the authors could be more specific when mentioning "reconfigurability" and provide more explanations if possible.  
+
+<|ref|>text<|/ref|><|det|>[[115, 541, 882, 639]]<|/det|>
+RE: We thank the reviewer for the comment. The reconfigurability refers both the fast chirping and comb spacing switching and tuning. For fast chirping, we have demonstrated unprecedented chirping rate up to \(2.4 \times 10^{17} \mathrm{~Hz} / \mathrm{s}\) orders of magnitude faster than other approaches. For comb spacing switching and tuning, we achieved it with high-speed EO modulation without involving complex tuning dynamics, which, again, is significantly faster than conventional Kerr microcombs. To make it clear, we have revised the related sentences in Abstract and Introduction.  
+
+<|ref|>text<|/ref|><|det|>[[115, 646, 882, 760]]<|/det|>
+4. The linewidth testing shown in Fig. 2e, which is obtained from a heterodyne measurement of the entire comb, is quite confusing to me. The authors claim that the tested linewidth is "the upper limit of the intrinsic linewidth of individual comb lines" (Page 4, line 222)? In my understanding, the noises of comb lines with higher signal powers would overwhelm those of low-power comb lines. As a result when normalizing the noise spectrum of the entire comb, one is probably seeing a larger noise contribution from higher-power lines, which are not necessarily better or worse than the lower-power lines. Please provide more references or proof to support this claim.  
+
+<|ref|>text<|/ref|><|det|>[[115, 766, 881, 799]]<|/det|>
+RE: We thank the reviewer for raising this point. The reviewer is correct and we have revised the relevant sentences in the paper. We provide more detailed linewidth characterizations in the following.  
+
+<|ref|>text<|/ref|><|det|>[[115, 805, 882, 886]]<|/det|>
+5. Following up on the above question, it would be really nice if the authors could characterize the linewidths of each comb line. The intrinsic linewidth variation among different comb lines could assist the understanding of the microcomb dynamics. For Kerr microcombs and EO combs, they both show the increasing linewidths when moving from the center to sides (Ref. [1, 2] below). I am curious about the linewidth variation trend of the proposed microcomb under different states as shown in Fig. 2b and
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[144, 90, 883, 123]]<|/det|>
+Fig. 5d. Specifically, for the ‘clustered’ comb state shown in Fig. 2b, Fig. S3b and d, would the linewidth increase from respective centers of different clusters?  
+
+<|ref|>text<|/ref|><|det|>[[143, 129, 883, 163]]<|/det|>
+[1] Lei, F., Ye, Z., Helgason, Ó. B., Fülöp, A., Girardi, M., & Torres- Company, V. (2022). Optical linewidth of soliton microcombs. Nature Communications, 13(1), 3161.  
+
+<|ref|>text<|/ref|><|det|>[[144, 168, 882, 217]]<|/det|>
+[2] Skehan, J. C., Naveau, C., Schroder, J., & Andrekson, P. (2021). Widely tunable, low linewidth, and high power laser source using an electro-optic comb and injection- locked slave laser array. Optics Express, 29(11), 17077- 17086.  
+
+<|ref|>text<|/ref|><|det|>[[115, 224, 882, 289]]<|/det|>
+RE: We thank the reviewer for the suggestion. We have performed detailed characterizations on the linewidths of individual comb lines. We used the setup shown in Fig. R1 below for the measurements, in which an individual comb line is separated from the rest of the comb by a tunable narrow- band fiber Bragg grating (FBG) filter and its linewidth is measured by the self- heterodyning method.  
+
+<|ref|>image<|/ref|><|det|>[[300, 319, 712, 456]]<|/det|>
+<|ref|>image_caption<|/ref|><|det|>[[130, 469, 866, 501]]<|/det|>
+<center>Figure R1 Schematic of the experimental setup used to characterize the linewidths of individual comb lines. </center>  
+
+<|ref|>text<|/ref|><|det|>[[115, 507, 882, 589]]<|/det|>
+Figure R2a shows the overall frequency noise spectra for different lasing states. For the comb states, the linewidth measurement was performed on the entire combs similar to Fig. 2 in the paper. It shows an overall linewidth of about \(7\mathrm{kHz}\) , \(7\mathrm{kHz}\) , and \(1.5\mathrm{kHz}\) , respectively, for the single- mode lasing state, the single- lobe comb state, and the bifurcated two- lobe comb state. These values are slightly worse than those shown in the paper simply due to the slight degradation of the LN device and the RSOA in the past four months.  
+
+<|ref|>text<|/ref|><|det|>[[115, 595, 882, 708]]<|/det|>
+The recorded linewidths for individual comb lines are shown in Fig. R2c and d for the two comb states. Due to the limited powers of the individual comb lines, we can only perform linewidth characterizations for the central portion of the combs. Figure R2c shows that, for the single- lobe comb state, the individual comb lines exhibit linewidths in the range of \(1 - 3\mathrm{kHz}\) . Figure R2d shows that one lobe of the comb exhibits individual comb linewidths in the range of \(0.8 - 2\mathrm{kHz}\) , while the other lobe exhibits slightly larger linewidths in the range of \(5 - 8\mathrm{kHz}\) . The exact physical reason for this difference is not clear at this moment, which requires further exploration.  
+
+<|ref|>text<|/ref|><|det|>[[115, 715, 882, 812]]<|/det|>
+Overall, the linewidths of individual comb lines do not show a certain deterministic trend across the comb spectrum, distinctive to the ones shown in Ref. [1,2] pointed out by the reviewer. This is understandable since the linewidth feature shown in Ref.[1] is due to the recoil effect between Raman- induced self- frequency shift and dispersive wave generation, and that shown in Ref.[2] is due to the noise multiplication during the cascaded sideband creation in a pure EO comb. The mechanism of comb generation is fundamentally different in our comb laser which could lead to different linewidth behaviors.
+
+<--- Page Split --->
+<|ref|>image<|/ref|><|det|>[[171, 92, 826, 466]]<|/det|>
+<|ref|>image_caption<|/ref|><|det|>[[114, 475, 883, 574]]<|/det|>
+<center>Figure R2 a) Frequency noise spectrum of a single-mode lasing state, single-lobe comb state (Comb 1), and bifurcated two-lobe comb state (Comb 2). For the combs, the frequency noise is for the entire combs, similar to Fig.2 in the paper. b) Optical spectrum (blue) and recorded linewidth (red dot) for the single-mode lasing state. c) Optical spectrum (blue) and recorded linewidths (red dots) of individual comb lines, for the single-lobe comb state (Comb 1). d) Same as c) but for the bifurcated two-lobe comb state (Comb 2). In b)-d), the small spectral gap on the noise floor is due to the spectral filtering by the FBG filter. </center>  
+
+<|ref|>text<|/ref|><|det|>[[115, 600, 882, 715]]<|/det|>
+6. The optical spectra presented in the experiment and simulation exhibit notable differences. Could the authors comment on the main reasons for the difference? Would the mismatch between the main laser cavity and the racetrack resonator influence the comb state? Will the comb power keeping increasing with higher RF power as claimed in Page 4 line 234? As the authors emphasize that the cavities are dispersion-engineered, what is the influence of the dispersion of different cavities (the racetrack cavity and the main laser cavity) on the comb states? Comprehensive analyses are needed to get a better understanding of the proposed microcomb.  
+
+<|ref|>text<|/ref|><|det|>[[115, 720, 880, 739]]<|/det|>
+RE: We thank the reviewer for the detailed questions. We answer the questions separately in the following:  
+
+<|ref|>text<|/ref|><|det|>[[115, 745, 618, 762]]<|/det|>
+I) Could the authors comment on the main reasons for the difference?  
+
+<|ref|>text<|/ref|><|det|>[[115, 769, 882, 850]]<|/det|>
+In fact, we are able to generate a comb from Laser \(\beta\) similar to the simulation, as we showed in Fig.S2 of the supplementary information. The mechanism underlying the bifurcated comb produced by Laser \(\alpha\) is currently still under investigation, which may result from complex physics not included in the simplified model. We have added one sentence in Section III.B of the supplementary information to make this point clear.  
+
+<|ref|>text<|/ref|><|det|>[[113, 857, 880, 875]]<|/det|>
+II) Would the mismatch between the main laser cavity and the racetrack resonator influence the comb state?
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 89, 883, 138]]<|/det|>
+Yes, the mismatch has strong impact on the stability and shape of the comb state. Chaotic comb may occur in a mismatched laser. We are currently still investigating the dynamics of the laser at different cavity length offset.  
+
+<|ref|>text<|/ref|><|det|>[[115, 144, 825, 162]]<|/det|>
+III) Will the comb power keeping increasing with higher RF power as claimed in Page 4 line 234?  
+
+<|ref|>text<|/ref|><|det|>[[115, 168, 883, 218]]<|/det|>
+The detailed relationship between the produced comb laser power and the RF driving power is shown in Fig. S3 a,c,f,h of the supplementary information for the two lasers. We observed comb power saturation in Laser \(\beta\) . In Laser \(\alpha\) , however, we haven't observed saturation within the applicable RF power range.  
+
+<|ref|>text<|/ref|><|det|>[[115, 225, 883, 257]]<|/det|>
+IV) As the authors emphasize that the cavities are dispersion-engineered, what is the influence of the dispersion of different cavities (the racetrack cavity and the main laser cavity) on the comb states?  
+
+<|ref|>text<|/ref|><|det|>[[115, 264, 883, 313]]<|/det|>
+Dispersion of the racetrack microresonator is crucial for the comb generation and mode locking. A slight anomalous dispersion is required, as we showed in the numerical modeling. The dispersion of the main laser cavity plays a minor role. We engineered it to be close to zero.  
+
+<|ref|>text<|/ref|><|det|>[[115, 320, 883, 368]]<|/det|>
+7. In Fig.2, the hybrid microcomb produces clustered microcombs, which are also shown in Fig. S3, while the spectral shapes in Fig. 5 are quite different. Could the authors comment on the possible origins of this difference. Is this due to the use of an optoelectronic feedback loop?  
+
+<|ref|>text<|/ref|><|det|>[[115, 374, 883, 439]]<|/det|>
+RE: We thank the reviewer for the question. We do observe different comb spectra for RF driven comb and feedback lock comb. Currently, the feedback locked comb is wider in spectrum. This could be related to that the dynamically adjusted feedbacked RF modulation assist in the generation of broader comb, but the detailed mechanism require further investigation in future works.  
+
+<|ref|>text<|/ref|><|det|>[[115, 446, 883, 495]]<|/det|>
+8. In Fig. 5, the authors show the feedback mode-locking of the comb laser. The concept is similar to the combination of EO combs and optoelectronic oscillators as demonstrated in previous works, such as Ref. [3-4] below. It is recommended to reference these works.  
+
+<|ref|>text<|/ref|><|det|>[[140, 502, 883, 535]]<|/det|>
+[3] Peng, H., Lei, P., Xie, X., & Chen, Z. (2021). Dynamics and timing-jitter of regenerative RF feedback assisted resonant electro-optic frequency comb. Optics Express, 29(26), 42435-42456.  
+
+<|ref|>text<|/ref|><|det|>[[140, 541, 883, 574]]<|/det|>
+[4] Sakamoto, T., Kawanishi, T., & Izutsu, M. (2006). Optoelectronic oscillator using a LiNbO3 phase modulator for self-oscillating frequency comb generation. Optics letters, 31(6), 811-813.  
+
+<|ref|>text<|/ref|><|det|>[[115, 581, 883, 614]]<|/det|>
+RE: We thank the reviewer for the suggestion. We have added these two references as Ref.[47, 48] in the paper.  
+
+<|ref|>text<|/ref|><|det|>[[115, 620, 883, 717]]<|/det|>
+9. There are several typos in describing Fig. 4: Page 5, line 290-291: "Fig. 4f-h show the temporal variation of the 15-GHz beating signal..." While the temporal variations are presented in Fig. 4b-d. Page 5, line 307-309: "The frequency tuning range of 1.2 GHz at the modulation speed of 100 MHz (Fig. 4h) corresponds to a frequency tuning rate..." The fast-chirping results are given in Fig. 4d. Page 5, line 314-315: "As shown in Fig. 4e, the device exhibits a frequency tuning efficiency..." The efficiency is shown in Fig. 4h.  
+
+<|ref|>text<|/ref|><|det|>[[115, 723, 757, 740]]<|/det|>
+RE: We thank the reviewer for finding these typos. We have corrected them in the paper.
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[115, 90, 303, 106]]<|/det|>
+## REVIEWERS' COMMENTS  
+
+<|ref|>text<|/ref|><|det|>[[115, 127, 393, 143]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 163, 740, 180]]<|/det|>
+The authors have addressed all my comments and I suggest publication for this paper.  
+
+<|ref|>text<|/ref|><|det|>[[115, 218, 393, 234]]<|/det|>
+Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 254, 882, 326]]<|/det|>
+The authors have addressed most of my comments. There is still a lot of physics here unexplored and not yet fully understood but I think it is reasonable to leave them to future studies as the results here are quite interesting and deserve to be published. There is one typo though: The unit of the phase noise in Fig. 5g should be "dBc/Hz".
+
+<--- Page Split --->

peer_reviews/126b45d3c59304c8cdcaeff2fb15da92849086ba5d613f1f0d668d95576aa34f/supplementary_0_Peer Review File/images_list.json

@@ -0,0 +1 @@
+[]

peer_reviews/126b45d3c59304c8cdcaeff2fb15da92849086ba5d613f1f0d668d95576aa34f/supplementary_0_Peer Review File/supplementary_0_Peer Review File.mmd

@@ -0,0 +1,429 @@
+
+# nature portfolio  
+
+# Peer Review File  
+
+T- Cell Dysfunction in the Glioblastoma Microenvironment is Mediated by Myeloid Cells Releasing Interleukin- 10  
+
+![PLACEHOLDER_0_0]
+  
+
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+
+## REVIEWER COMMENTS  
+
+Reviewer #1 (Remarks to the Author):  
+
+This manuscript studies the crosstalk between infiltrating microglia/macrophages and T- cells in the immunosuppressive microenvironment of glioblastoma. This is an important topic that can offer further insights into the mechanisms of resistance to anti- PD1 immunotherapy (Zhao et al. Nat. Medicine 25 (2019)). This work is a continuation of previously published work (Herik Heiland et al., Nat. Commun. 10 (2019)), where the authors found that the crosstalk between microglia cells and reactive astrocytes is responsible for upregulating IL- 10 release in glioblastoma through JAK/STAT signaling. In this manuscript, the authors show that IL- 10 secreted by HMOX1+ myeloid cells is responsible for inducing a dysfunctional state in T cells infiltrating mesenchymal regions of the tumor.  
+
+There are several aspects that in my opinion need to be addressed before publication:  
+
+1. Some of the statements in the Abstract, Introduction, and Discussion do not seem to have clear support in the data and analyses that are presented in the manuscript. Either those statements need to be more carefully crafted or the data and analyses supporting them need to be more clearly presented:  
+
+1a. Lines 375-378: the authors state that their analysis indicates that the dysfunctional state of T cells appears to be a transient state. However, this reviewer was unable to find any data or analyses supporting the transitory character of dysfunctional T cell states in subsection "Dysfunctional State of T cells is Driven by IL-10 Signaling".  
+
+1b. Lines 58, 117-119, and 394-396: the authors mention that HMOX1+ myeloid cells co-localize spatially with the mesenchymal signature of glioblastoma. However, Figs. 3f,g only show the spotwise correlation between dysfunctional T cell gene expression markers (HAVCR2 and LAG3) and the mesenchymal gene expression signature, but as far as I see they do not present any analysis of the spot-wise correlation between HMOX1+ myeloid cell markers and the mesenchymal gene expression signature. Similarly, in subsection "T cell Activation and Exhaustion Reveals Spatial Heterogeneity and Association with Glioblastoma Subtypes" there seem to be no results about the spatial pattern of HMOX1+ myeloid cells.  
+
+1c. Lines 403-405. The authors state that their experiments with the ex vivo neocortical GBM model confirm that HMOX1+ myeloid cells cause a reduction of effector T cells. However, in these experiments both HMOX1+ and HMOX1- myeloid cells are depleted using clodronate. In absence of other data, these experiments only show that myeloid cells cause a reduction of effector T cells.  
+
+1d. Lines 352-356 and 414-418, and Fig 4q-r. It is unclear what comparison was performed here. How did the authors determine that there is a significant enrichment for activated T cells and B cells upon JAK/STAT inhibition in the patient? A more detailed explanation of the comparison and the assumptions that were made would be useful here.  
+
+2. The authors introduce the "nearest functionally connected neighbor" algorithm to infer candidate paracrine interactions from the single-cell RNA-seq data. Since the performance of this algorithm has not been rigorously evaluated, it is hard to know how reliable the results of this method are in general. In Supplementary Fig. 5, the authors show that the more conventional and established algorithm NicheNet (Browaeys et al. Nat. Methods 17 (2020)) also finds the candidate interaction between myeloid cells (expressing IL10) and T-cells (expressing IL10RA). However, NicheNet does not show that this interaction is specific to HOMX1+ myeloid cells. Can the authors present any other evidence from the single-cell RNA-seq data to support their statement that HOMX1+ cells are mostly responsible for this interaction (e.g. correlation between HOMX1+ and IL10 expression within myeloid cells)? It would be also useful to show the UMAP representation labeled by HMOX1 expression.  
+
+3. The description of the "nearest functionally connected neighbor" algorithm in the Methods section
+
+<--- Page Split --->
+
+lacks much technical detail. It would be useful to include details about the models, fitting methods, etc., and rewrite the description of the algorithm more carefully.  
+
+4. I found the main figures to be unnecessarily complex with 10-18 panels each. The authors might consider keeping the panels that convey the main results and moving the rest of the panels to supplementary figures. There are also several typos that need to be corrected. For example, the panels in Fig. 3 are mismatched with the figure legend (e.g. 3c and 3d seem to be exchanged) and with the main text (lines 283-310). Supplementary Fig. 6 includes a caption "Supplementary Fig. 5" which should be removed. The legend of Supplementary Fig. 3 says "Dimensional reduction (UMAP) of gene expression of the different simulation experiments". However, what is shown in the figure seems to be the UMAP of the single-cell RNA-seq data from the patients colored by the imputed stimulation signatures.  
+
+Reviewer #2 (Remarks to the Author):  
+
+NCOMMS- 21- 07876- T Crosstalk between lymphoid and myeloid cells orchestrates glioblastoma immunity through Interleukin 10 signaling  
+
+In their manuscript titled "Crosstalk between lymphoid and myeloid cells orchestrates glioblastoma immunity through Interleukin 10 signaling", investigators leveraged single- cell and spatial transcriptomics to infer cellular crosstalk between macrophages/microglia and T cells within human GBM samples. From eight GBM samples, 21 clusters were identified, where many of these clusters represented macrophages and microglia. Sub- clustering of the T cell cluster revealed different activation states that were then mapped with pseudo- time and RNA velocity analyses. This provided a differentiation map from naive to terminally exhausted, with an intermediate highly proliferative state. Additionally, by combining these techniques, authors identified an association between the mesenchymal GBM subtype and T cell exhaustion. Furthermore, authors developed a new model, nearest functional connected neighbor, to identify ligand/receptor interactions from scRNAseq data. Computational analyses are thorough and novel; however, the follow- up mechanistic studies, using an in vitro neocortical model, were limited in their support of computational findings. Primary concerns are related to the impact of the mechanistic studies. These concerns are detailed below:  
+
+1. A major limitation is the in vitro neocortical model used in Figure 4. Authors state that this model replicates the tumor microenvironment because it is derived from brain tissue; however, during tumorigenesis the TME is largely shaped by the infiltrating immune cells (1), which are absent in this model. Authors should comment on the "myeloid" cells that exist within the neocortical tissue that are being evaluated in figure 4, which should primarily be microglia and not the HMOX1+ macrophages from the computational studies. Along these lines, a quantification of tumor growth in the myeloid depleted condition is important, as myeloid cells generally promote tumor growth, so their absence alone may have an effect on tumor growth independent of T cells or IL10 inhibition.  
+
+2. Additionally, the short incubation time (3 days) from tumor injection and subsequent T cell transfer may not provide sufficient time for meaningful interactions and subsequent functional outputs to occur, for example T cell exhaustion. Regarding the transferred T cells, it was not clearly stated whether there is selection for tumor-specific T cells from patient blood, therefore T cells that are injected into the neocortical model may or may not react to the tumor. Evidence of T cell recognition of tumor cells through killing or activation is necessary to increase the impact of this model. Much of T cell exhaustion biology is ignored in this model, such as the conditions and locations under which priming occur and trafficking to the tumor site. Therefore, conclusions that can be drawn from this model are limited. Furthermore, authors should address potential allogeneic reactivity to the cell line BTSC#233 by patient T cells.  
+
+3. Although important for spatial information, the immunofluorescence images shown without
+
+<--- Page Split --->
+
+quantifications in figure 4 are not sufficient to validate the computational studies. For example, in figure 4g, TIM3 is used to identify "exhausted T cells", but it has been shown that TIM3 can also be a marker of terminal effector differentiation. Therefore, this would be more convincing if other parameters were used to identify this population, such as PD1 or other functional studies showing T cell activity.  
+
+4. When looking at the contribution of different patients to the final clusters, it is apparent that many clusters are specific to single patients. In particular, \(80\%\) of the HMOX1+ group is made up of a single patient. What impact does this have on the broader applicability of these findings? Would this bias for a single patient carry over into the downstream analyses? Is this to be expected whenever pooling human samples?  
+
+5. The approval for the use of an un-licensed drug in a GBM patient needs to be specifically addressed within the ethics section.  
+
+## 6. Minor concerns:  
+
+a. The conclusion that myeloid and lymphoid interactions lead to T cell dysfunction through IL10 is vague and not unoriginal. Secretion of IL10 by myeloid cells is not a novel finding, nor is the role of IL10 in T cell dysfunction(2,3).  
+b. Using in vitro stimulated T cells to compare with in vivo T cells coming from a tumor ignores much of the complexity in signals that is occurring intratumorally.  
+c. Using a second cell line in the in vitro neocortical studies would increase the impact of related findings.  
+
+In general, authors show novel and interesting computational analyses from cutting edge techniques, however they lack substance in their follow up mechanistic studies. The use of pseudotime and RNA velocity to interrogate T cell activation states and pair them with GBM subtypes and spatial information are intriguing. Additionally, the nearest functionally connected neighbor algorithm will add to the expanding pool of resources for inferring intercellular communication from scRNAseq data. Follow-up studies are limited in their support of computational findings (IL- 10 mediated T cell exhaustion); therefore, additional validation studies are required, or the focus of the story should shift to highlight the novel bioinformatic analyses.  
+
+## Reference:  
+
+1. Salmon, H., Remark, R., Gnjatic, S. and Merad, M., 2019. Host tissue determinants of tumour immunity. Nature Reviews Cancer, 19(4), pp.215-227.  
+2. McLane, L.M., Abdel-Hakeem, M.S. and Wherry, E.J., 2019. CD8 T cell exhaustion during chronic viral infection and cancer. Annual review of immunology, 37, pp.457-495.  
+3. Quail, D.F. and Joyce, J.A., 2017. The microenvironmental landscape of brain tumors. Cancer cell, 31(3), pp.326-341.  
+
+Reviewer #3 (Remarks to the Author):  
+
+Ravi, Neidert, Will et al present a single- cell RNA- sequencing study of tumor- infiltrating lymphocytes of patients with GBM. The profile 8 patients and additionally show data for 3 additional patients using spatial profiling, all using 10x platforms. Using established and novel analytical tools to infer trajectories of cell differentiation, they identify variability among T cells; among other, they find subclusters of CD8+ T cells with high expression of dysfunction marker TIM3 (HAVCR2) and cells with a hypoxia signature with distinct trajectories. They correlate these signatures with signatures of T cells collected following in vitro stimulation with different cytokines, including IL2, IFNG and IL10, arguing that IL10 stimulated cells have a lower activation score compared to other cells identified as effector cells. Using spatial RNA- seq of three tumors, they find an association of tumor- mesenchymal niches
+
+<--- Page Split --->
+
+and infiltration of exhausted T cells (TIM3/LAG3), suggesting that niches of dysfunctional T cells may be in part explained by cancer cell intrinsic features. In order to understand the origin of IL10 and to solidify the role of T cells as IL10 recipients, they present a analytical framework that infers ligandreceptor interactions using several constraints, and demonstrate that myeloid cells (CD163+, HMOX1+) are a main source of IL10. To begin validating this finding, they use slice cultures that they deplete of myeloid cells and show that depletion of myeloid cells results in reduction of IL10 (in the presence of tumor cells); in these models, they then co- culture autologous T cells and show that depletion of myeloid cells in slice cultures results in increased IL2, but not IFNG protein abundance. Incubation of T cells with an IL10R- inhibitor prior to co- culture with tissues results in increased IL2 production in T cells. Because the JAK/STAT pathway is downstream of IL10 signaling, they use ruxolitinib (a selective JAK1/2 inhibitor) first i slice model showing increased IL2, and then use this drug in the neo- adjuvant therapy of a patient with GBM, followed by analysis of the surgical specimen, which shows activation of T cells.  
+
+GBM is a disease with extremely poor prognosis, and therapeutic development has in part been hampered by limited understanding of the tumor microenvironment; as such, the study of potentially high importance. However, several aspects raised significant concerns and reduced enthusiasm for this study, and need to be addressed,  
+
+Major points:  
+
+1. Nowhere in the manuscript do the authors describe the characteristics of the patient tumors used for either single-cell sequencing of spatial sequencing. Are these all treatment naive tumors? where they exposed to different therapies (radiation, chemotherapy, immunotherapy, investigational drugs) - this will have a dramatic impact on the measured T cell phenotypes and in and of itself could describe variability seen in the data set. The authors should describe basic demographics and treatment history; it is not reasonable to request from the authors to attempt to account for variability based on basic demographics (this, and virtually any single-cell study would be underpowered), but they should show major analyses/findings in the context of different therapies received to exclude the possibility noted above.  
+
+2. Technical quality: it is somewhat surprising that the authors only recover \(\sim 1000\) unique genes per cell with only \(\sim 2300\) unique molecular identifiers - this is not on par with the quality described and raises concerns regarding data quality; this is particularly surprising as they used the 3.1 chemistry which performs better than prior chemistries; in fact recent studies performing profiling from frozen tissues even achieved similar or better quality compared to this present study (Slyper, Nature Medicine, 2020). Furthermore, some of the clusters described might be artifactual due to tissue processing (e.g. "hypoxia cluster")- this possibility should be addressed using available data sets systematically investigating such artifacts (e.g. Ido Amit laboratory). Furthermore, the authors should comment on the technical quality. An additional embedding showing the UML count for major analyses should be shown in the supplement to exclude the possibility of technical artifacts as drivers of clustering.  
+
+3. There is no statistical evaluation of the inference made in Figure 3f - the authors should provide this; in this same figure, they also show that CCL2 myeloid cells are scoring highly, which is a gene considered to be an immunostimulatory gene/protein - how do they reconcile this? This brings up the question about a more nuanced annotation of the myeloid cells beyond monocytes, macrophages and microglia. The effect size of the gene set enrichment analysis in 3h is very underwhelming. In fact, throughout this section and the studies shown in Figure 4, the effect sizes are very small with in part borderline significance, raising the question of biological significance of these findings.  
+
+4. The experiments in slice cultures should be described in more detail in the main text. Here, they state that they performed "myeloid depletion" when in fact they performed microglia depletion (as stated as header in the methods section). The effect size of myeloid/microglia depletion on IL10 production is rather modest. One missing control is depletion of other cell types within the slice culture and measurement of the effect on IL10 to exclude the possibility that there are other major sources of
+
+<--- Page Split --->
+
+IL10 production (which is likely). Furthermore, it is surprising that IL2 production, but not IFNG production increases during T cell depletion - the authors should offer potential explanations as this argues against reinvigoration of T cell poly- functionality. The results shown in 4j confirm that IL10 is an immunosuppressive cytokine, but do not substantiate claims that this is medicated by myeloid cells. Again, perplexing that no change in interferon gamma is seen. The single patient study is encouraging - was this pre- /post- comparison performed after single- agent therapy with ruxolitinib or was there a combination used? If the latter, it is possible that observed effects are due to other treatment constituents (see comment 1).
+
+<--- Page Split --->
+
+## Point - by - Point  
+
+D. H. Heiland  
+
+We thank all reviewers for their time and effort in evaluating our manuscript and appreciate the positive feedback on our project. We have tried to mitigate the issues highlighted by the reviewers, which has led to a significant improvement in the quality of our manuscript. The following main changes have been made in this context:  
+
+## 1. Quality of the scRNA-seq experiments.  
+
+By resequencing the libraries, the quality of the entire dataset was significantly improved and the number of detected genes as well as the UMIs per cell were significantly increased.  
+
+## 2. Analysis and data integration  
+
+By applying more advanced algorithms for horizonal and vertical data integration as well as cell type alignment, we were able to present a clearer picture of T cell diversity in the tumors. We separated CD4 and CD8 positive T cells for all downstream analysis.  
+
+## 3. Avoid overfitting in the NFCN algorithm  
+
+Our cell communication algorithm has been optimized to reduce potential overfitting and improve prediction. For this purpose, we integrated multiple prediction/validation layers and external algorithms. The new version is also compatible with the conventional scRNA- seq tools (Seurat) and available as an R package (NFCN2).  
+
+## 4. Structure of the manuscript and presentation  
+
+We restructured our manuscript to present clear hypothesis- driven argumentation and pointed out limitations and ambiguities. The illustrations have been simplified to improve general understanding.  
+
+## 5. Validation model  
+
+Our experimental model is not without limitations, which are discussed in detail. We performed new experiments and analysis to improve the experimental validation.  
+
+## 6. Clinical Dataset  
+
+Our in- vivo dataset is now described in detail, and we were able to generate further data to strengthen our hypothesis.  
+
+In order to discuss the reviewer comments in detail, we provide a point- by- point discussion.
+
+<--- Page Split --->
+
+## Reviewer #1 (Remarks to the Author):  
+
+This manuscript studies the crosstalk between infiltrating microglia/macrophages and T- cells in the immunosuppressive microenvironment of glioblastoma. This is an important topic that can offer further insights into the mechanisms of resistance to anti- PD1 immunotherapy (Zhao et al. Nat. Medicine 25 (2019)). This work is a continuation of previously published work (Henrik Heiland et al., Nat. Commun. 10 (2019)), where the authors found that the crosstalk between microglia cells and reactive astrocytes is responsible for upregulating IL- 10 release in glioblastoma through JAK/STAT signaling. In this manuscript, the authors show that IL- 10 secreted by HMOX1+ myeloid cells is responsible for inducing a dysfunctional state in T cells infiltrating mesenchymal regions of the tumor.  
+
+We would like to thank the reviewer for his time and comments leading to an improvement of the manuscript.  
+
+There are several aspects that in my opinion need to be addressed before publication:  
+
+1. Some of the statements in the Abstract, Introduction, and Discussion do not seem to have clear support in the data and analyses that are presented in the manuscript. Either those statements need to be more carefully crafted or the data and analyses supporting them need to be more clearly presented:  
+
+We have substantially revised our argumentation to better support the data presented and to clearly define our hypotheses. By improving the scRNA-seq datasets and analysis approaches, some of our previously stated hypotheses have been relativized. The major difference compared to our previous manuscript is a separation of CD8 and CD4 positive T cells for all further sub analysis. Our new data incorporated novel aspects of the underlying mechanism of tumor-associated T cell response.  
+
+1a. Lines 375- 378: the authors state that their analysis indicates that the dysfunctional state of T cells appears to be a transient state. However, this reviewer was unable to find any data or analyses supporting the transitory character of dysfunctional T cell states in subsection "Dysfunctional State of T cells is Driven by IL- 10 Signaling".  
+
+In our revised version, we performed model integration of RNA- velocity and lineage tree reconstruction to improve the exploration of state specific pathway activation. We found that IL10 response was highly correlated with the expression of exhaustion programs in two T cell clusters. We rewrote this part of the manuscript to improve understanding and removed statements that are no longer supported or have caused confusion.  
+
+1b. Lines 58, 117- 119, and 394- 396: the authors mention that HMOX1+ myeloid cells co- localize spatially with the mesenchymal signature of glioblastoma. However, Figs. 3f,g only show the spot- wise correlation between dysfunctional T cell gene expression markers (HAVCR2 and LAG3) and the mesenchymal gene expression signature, but as far as I see they do not present any analysis of the
+
+<--- Page Split --->
+
+spot- wise correlation between HMOX1+ myeloid cell markers and the mesenchymal gene expression signature. Similarly, in subsection "T cell Activation and Exhaustion Reveals Spatial Heterogeneity and Association with Glioblastoma Subtypes" there seem to be no results about the spatial pattern of HMOX1+ myeloid cells.  
+
+We thank the reviewer for picking up on this lack of clarity in presentation. We have improved the text and figure legends for clarity. In our revised version of the manuscript, we added spot- wise correlations to support our hypothesis as well as spatial data analysis of the model.  
+
+1c. Lines 403- 405. The authors state that their experiments with the ex vivo neocortical GBM model confirm that HMOX1+ myeloid cells cause a reduction of effector T cells. However, in these experiments both HMOX1+ and HMOX1- myeloid cells are depleted using clotronate. In absence of other data, these experiments only show that myeloid cells cause a reduction of effector T cells.  
+
+Thank you for this helpful comment. It is indeed the case that we remove all myeloid cells and therefore are unable to differentiate HMOX1 pos/neg myeloid cells individually. Using our human model, we are currently not able to specifically target HMOX1 positive cells. We have described these limitations. To approach this limitation, we quantified the spatial distance of HMOX1- positive and - negative cells in our model and concluded that HMOX- positive cells are mainly localized in the proximity of the tumor. Thus, we assumed that HMOX1- negative cells were only present to a small extent within the tumor. However, in our more detailed analysis of the slice model, we discuss limitations and cofounders more detailed.  
+
+1d. Lines 352- 356 and 414- 418, and Fig 4q- r. It is unclear what comparison was performed here. How did the authors determine that there is a significant enrichment for activated T cells and B cells upon JAK/STAT inhibition in the patient? A more detailed explanation of the comparison and the assumptions that were made would be useful here.  
+
+This part was fully rewritten for an improved presentation of our hypothesis. The data are re- analyzed in accordance with our new findings.  
+
+2. The authors introduce the "nearest functionally connected neighbor" algorithm to infer candidate paracrine interactions from the single-cell RNA-seq data. Since the performance of this algorithm has not been rigorously evaluated, it is hard to know how reliable the results of this method are in general. In Supplementary Fig. 5, the authors show that the more conventional and established algorithm NicheNet (Browaeys et al. Nat. Methods 17 (2020)) also finds the candidate interaction between myeloid cells (expressing IL10) and T-cells (expressing IL10RA). However, NicheNet does not show that this interaction is specific to HOMX1+ myeloid cells. Can the authors present any other evidence from the single-cell RNA-seq data to support their statement that HOMX1+ cells are mostly responsible
+
+<--- Page Split --->
+
+for this interaction (e.g. correlation between HOMX1+ and IL10 expression within myeloid cells)? It would be also useful to show the UMAP representation labeled by HMOX1 expression.  
+
+In the new version of our "nearest functionally connected neighbor" (NFCN) algorithm, we implemented various new functions. In general, NFCN is built to quantify cellular interactions of a defined pathway (in our case the IL10- IL10R interaction). In comparison to NicheNet and CellChat, we inferred cellular connections based on the likelihood of cell pairs from the scRNA- seq dataset. Indeed, this quantification leads to overfitting as long as the ground truth is unknown. In order to overcome this problem, we redesigned the algorithm to integrate 3 data layers for improved prediction of cellular interactions.  
+
+1. Prediction of the cell-pair likelihood based on scRNA-seq data.  
+
+2. Deconvolution of Cell-Cell signaling from doublets  
+
+3. Integration of spatial resolved transcriptomics to confirm spatial juxta positioning of cell pairs.  
+
+We further integrated an unsupervised model using CellChat to infer the most common Cell- Cell interaction across clusters. Through our optimization, we tailored the model to predict cellular communication and reduced bias. We have added a supplementary result part to explain this model in detail.  
+
+3. The description of the "nearest functionally connected neighbor" algorithm in the Methods section lacks much technical detail. It would be useful to include details about the models, fitting methods, etc., and rewrite the description of the algorithm more carefully.  
+
+As mentioned in the answer above, we added supplementary results with detailed information.  
+
+4. I found the main figures to be unnecessarily complex with 10-18 panels each. The authors might consider keeping the panels that convey the main results and moving the rest of the panels to supplementary figures. There are also several typos that need to be corrected. For example, the panels in Fig. 3 are mismatched with the figure legend (e.g. 3c and 3d seem to be exchanged) and with the main text (lines 283-310). Supplementary Fig. 6 includes a caption "Supplementary Fig. 5" which should be removed. The legend of Supplementary Fig. 3 says "Dimensional reduction (UMAP) of gene expression of the different simulation experiments". However, what is shown in the figure seems to be the UMAP of the single-cell RNA-seq data from the patients colored by the imputed stimulation signatures.  
+
+Thanks for pointing out the typos and complexity of the figures. We have adapted the figures to facilitate ease of understanding.
+
+<--- Page Split --->
+
+## Reviewer #2 (Remarks to the Author):  
+
+Crosstalk between lymphoid and myeloid cells orchestrates glioblastoma immunity through Interleukin 10 signaling. In their manuscript titled "Crosstalk between lymphoid and myeloid cells orchestrates glioblastoma immunity through Interleukin 10 signaling", investigators leveraged single- cell and spatial transcriptomics to infer cellular crosstalk between macrophages/microglia and T cells within human GBM samples. From eight GBM samples, 21 clusters were identified, where many of these clusters represented macrophages and microglia. Sub- clustering of the T cell cluster revealed different activation states that were then mapped with pseudo- time and RNA velocity analyses. This provided a differentiation map from naive to terminally exhausted, with an intermediate highly proliferative state. Additionally, by combining these techniques, authors identified an association between the mesenchymal GBM subtype and T cell exhaustion. Furthermore, authors developed a new model, nearest functional connected neighbor, to identify ligand/receptor interactions from scRNAseq data. Computational analyses are thorough and novel; however, the follow- up mechanistic studies, using an in vitro neocortical model, were limited in their support of computational findings. Primary concerns are related to the impact of the mechanistic studies. These concerns are detailed below:  
+
+We would like to thank the reviewer for his time and comments leading to an improvement of the manuscript.  
+
+1. A major limitation is the in vitro neocortical model used in Figure 4. Authors state that this model replicates the tumor microenvironment because it is derived from brain tissue; however, during tumorigenesis the TME is largely shaped by the infiltrating immune cells (1), which are absent in this model. Authors should comment on the "myeloid" cells that exist within the neocortical tissue that are being evaluated in figure 4, which should primarily be microglia and not the HMOX1+ macrophages from the computational studies. Along these lines, a quantification of tumor growth in the myeloid depleted condition is important, as myeloid cells generally promote tumor growth, so their absence alone may have an effect on tumor growth independent of T cells or IL10 inhibition.  
+
+Thank you for this valuable comment. We have addressed and discussed this limitation in detail. There is no doubt that the myeloid cells within the presented model are predominantly composed of microglial cells. However, these cells can also transform reactively and consequently become HMOX1 positive. HMOX1 positive microglial cells also play a crucial role in other pathologies such as traumatic brain injury and subarachnoid hemorrhage. Therefore, it would be safe to assume that although the model is limited, the specific role associated with HMOX1 expression can be associated with activated microglial cells. We have discussed this limitation in detail.  
+
+Regarding the quantification of tumor growth in myeloid depletion condition: This question is of high interest and our laboratory is currently working on this interaction. At the moment, we feel that the addition of this data will result in a loss of focus of the results presented in this manuscript.
+
+<--- Page Split --->
+
+2. Additionally, the short incubation time (3 days) from tumor injection and subsequent T cell transfer may not provide sufficient time for meaningful interactions and subsequent functional outputs to occur, for example T cell exhaustion. Regarding the transferred T cells, it was not clearly stated whether there is selection for tumor-specific T cells from patient blood, therefore T cells that are injected into the neocortical model may or may not react to the tumor. Evidence of T cell recognition of tumor cells through killing or activation is necessary to increase the impact of this model. Much of T cell exhaustion biology is ignored in this model, such as the conditions and locations under which priming occur and trafficking to the tumor site. Therefore, conclusions that can be drawn from this model are limited. Furthermore, authors should address potential allogeneic reactivity to the cell line BTSC#233 by patient T cells.  
+
+Thank you for the detailed review of the model, which aids illustrating the various aspects, functionalities and limitations. Indeed, we are limited in the interpretation of our results. However, the following points deserve to be considered:  
+
+1. Regarding the first part of the question: We did not isolate tumor-specific T cells (mutation-associated neoantigens (MANA) associated TILs) from blood. In the context of brain tumors, to purify MANA-TILs is extremely challenging and only insufficiently possible using current methods. The aim of our model was to generate a T cell response and investigate the role of the tumor-associated microenvironment. Injection of a primary cell line which causes an allogeneic response is part of the model. Without this stimulus, a T cell response, as you mentioned above, is limited. This allogeneic reactivity should therefore be considered as intentional.  
+
+2) Regarding the second part of the question: Our data show that T cell activity (GZMB) in the tumor region becomes detectable after 3 days. (See data presented). The temporal dimensions of our slice model span a few days because the tumor infiltrates a large portion of the slice within 7 days. We have already reported tumor growth times in our previous publications<sup>1,2</sup>.  
+
+3. Although important for spatial information, the immunofluorescence images shown without quantifications in figure 4 are not sufficient to validate the computational studies. For example, in figure 4g, TIM3 is used to identify "exhausted T cells", but it has been shown that TIM3 can also be a marker of terminal effector differentiation. Therefore, this would be more convincing if other parameters were used to identify this population, such as PD1 or other functional studies showing T cell activity.  
+
+We added a more sophisticated validation of the imaging results. We choose Tim3 to confirm the results from the computational studies, in which the tissue resident memory cluster revealed the strongest enrichment of exhaustion markers.
+
+<--- Page Split --->
+
+4. When looking at the contribution of different patients to the final clusters, it is apparent that many clusters are specific to single patients. In particular, \(80\%\) of the HMOX1+ group is made up of a single patient. What impact does this have on the broader applicability of these findings? Would this bias for a single patient carry over into the downstream analyses? Is this to be expected whenever pooling human samples?  
+
+This problem was based on the vertical integration algorithm which has been fully revised.  
+
+5. The approval for the use of an un-licensed drug in a GBM patient needs to be specifically addressed within the ethics section.  
+
+The treatment was performed as part of the "Compassionate Use" program (RL 2001/83/EG VO 726/2004). We added explanations in the manuscript.  
+
+6. Minor concerns: a. The conclusion that myeloid and lymphoid interactions lead to T cell dysfunction through IL10 is vague and not unoriginal. Secretion of IL10 by myeloid cells is not a novel finding, nor is the role of IL10 in T cell dysfunction(2,3).  
+
+Indeed, this mechanism is reported in other cancer types but not for brain malignancy so far. Other cancer types can also be treated with checkpoint inhibitors, which is not possible for GBM. We think that investigating this special environment expands our comprehension. The fact that we found similar mechanism that can be also observed in other cancer types is not unexpected.  
+
+b. Using in vitro stimulated T cells to compare with in vivo T cells coming from a tumor ignores much of the complexity in signals that is occurring intratumorally.  
+
+The stimulation experiments are used to detect downstream pathway activation based on an isolated cytokine. We remove all other interpretations.  
+
+c. Using a second cell line in the in vitro neocortical studies would increase the impact of related findings.  
+
+Since this work does not focus on the tumor directly, using multiple donors to investigate the variance across patients was our main focus.  
+
+In general, authors show novel and interesting computational analyses from cutting edge techniques, however they lack substance in their follow up mechanistic studies. The use of pseudotime and RNA velocity to interrogate T cell activation states and pair them with GBM subtypes and spatial information are intriguing. Additionally, the nearest functionally connected neighbor algorithm will add to the expanding pool of resources for inferring intercellular communication from scRNAseq data. Follow- up
+
+<--- Page Split --->
+
+studies are limited in their support of computational findings (IL- 10 mediated T cell exhaustion); therefore, additional validation studies are required, or the focus of the story should shift to highlight the novel bioinformatic analyses.  
+
+Thank you for the appreciation of the computational results and tools. However, we think that biological validation, even if limited, supports the computational analysis.
+
+<--- Page Split --->
+
+Reviewer #3 (Remarks to the Author):  
+
+Reviewer #3 (Remarks to the Author):Ravi, Neidert, Will et al present a single- cell RNA- sequencing study of tumor- infiltrating lymphocytes of patients with GBM. The profile 8 patients and additionally show data for 3 additional patients using spatial profiling, all using 10x platforms. Using established and novel analytical tools to infer trajectories of cell differentiation, they identify variability among T cells; among other, they find sub- clusters of CD8+ T cells with high expression of dysfunction marker TIM3 (HAVCR2) and cells with a hypoxia signature with distinct trajectories. They correlate these signatures with signatures of T cells collected following in vitro stimulation with different cytokines, including IL2, IFNG and IL10, arguing that IL10 stimulated cells have a lower activation sore compared to other cells identified as effector cells. Using spatial RNA- seq of three tumors, they find an association of tumor- mesenchymal niches and infiltration of exhausted T cells (TIM3/LAG3), suggesting that niches of dysfunctional T cells may be in part explained by cancer cell intrinsic features. In order to understand the origin of IL10 and to solidify the role of T cells as IL10 recipients, they present a analytical framework that infers ligand- receptor interactions using several constraints, and demonstrate that myeloid cells (CD163+, HMOX1+) are a main source of IL10. To begin validating this finding, they use slice cultures that they deplete of myeloid cells and show that depletion of myeloid cells results in reduction of IL10 (in the presence of tumor cells); in these models, they then co- culture autologous T cells and show that depletion of myeloid cells in slice cultures results in increased IL2, but not IFNG protein abundance. Incubation of T cells with an IL10R- inhibitor prior to co- culture with tissues results in increased IL2 production in T cells. Because the JAK/STAT pathway is downstream of IL10 signaling, they use ruxolitinib (a selective JAK1/2 inhibitor) first slice model showing increased IL2, and then use this drug in the neo- adjuvant therapy of a patient with GBM, followed by analysis of the surgical specimen, which shows activation of T cells.  
+
+GBM is a disease with extremely poor prognosis, and therapeutic development has in part been hampered by limited understanding of the tumor microenvironment; as such, the study of potentially high importance. However, several aspects raised significant concerns and reduced enthusiasm for this study, and need to be addressed,  
+
+We would like to thank the reviewer for his time and comments leading to an improvement of the manuscript.  
+
+Major points:  
+
+1. Nowhere in the manuscript do the authors describe the characteristics of the patient tumors used for either single-cell sequencing of spatial sequencing. Are these all treatment naive tumors? where they exposed to different therapies (radiation, chemotherapy, immunotherapy, investigational drugs) - this will have a dramatic impact on the measured T cell phenotypes and in and of itself could describe variability seen in the data set.  
+
+All samples used for the dataset are naive non-treated primary GBM samples except the JAK-inhibitor treated samples as described in the last section on the manuscript.
+
+<--- Page Split --->
+
+The authors should describe basic demographics and treatment history; it is not reasonable to request from the authors to attempt to account for variability based on basic demographics (this, and virtually any single- cell study would be underpowered), but they should show major analyses/findings in the context of different therapies received to exclude the possibility noted above.  
+
+Indeed, prior treatment can strongly affect the immune compartment. Here, only primary non- treated samples are included. We have added a supplementary table for demographic details.  
+
+2. Technical quality: it is somewhat surprising that the authors only recover \(\sim 1000\) unique genes per cell with only \(\sim 2300\) unique molecular identifiers - this is not on par with the quality described and raises concerns regarding data quality; this is particularly surprising as they used the 3.1 chemistry which performs better than prior chemistries; in fact recent studies performing profiling from frozen tissues even achieved similar or better quality compared to this present study (Slyper, Nature Medicine, 2020). Indeed, the sequencing depth of the samples was only moderate (Sequencing Saturation \(\sim 10\% - 15\%\) ) and we decided to re-sequence all our libraries. We gained an improvement in quality to approximately \(\sim 2400\) genes per cell and \(\sim 10k\) unique molecular identifiers. To provide an overview of the dataset quality, we have added some comparisons to recent published datasets, Supplementary Figure 2.  
+
+Furthermore, some of the clusters described might be artifactual due to tissue processing (e.g. "hypoxia cluster")- this possibility should be addressed using available data sets systematically investigating such artifacts (e.g. Ido Amit laboratory). Furthermore, the authors should comment on the technical quality. An additional embedding showing the UMI count for major analyses should be shown in the supplement to exclude the possibility of technical artifacts as drivers of clustering.  
+
+We added a scRNA- seq quality check in the supplementary results. Using a recently reported cell- type alignment algorithm (WNN<sup>3</sup>), we redesigned the first part and specifically investigated CD8 and CD4 T cells separately. As recommended, we have opted for alignment to reference datasets. The stress cluster within the T cell population has already been confirmed in the meantime in another cohort<sup>4</sup>.  
+
+3. There is no statistical evaluation of the inference made in Figure 3f - the authors should provide this; in this same figure, they also show that CCL2 myeloid cells are scoring highly, which is a gene considered to be an immunostimulatory gene/protein - how do they reconcile this? This brings up the question about a more nuanced annotation of the myeloid cells beyond monocytes, macrophages and microglia. The effect size of the gene set enrichment analysis in 3h is very underwhelming. In fact, throughout this section and the studies shown in Figure 4, the effect sizes are very small with in part borderline significance, raising the question of biological significance of these findings.  
+
+We have revised large parts of the results presented formerly in Figure 4. This was done by resequencing and improving the vertical integration, have eliminated the previously seen myeloid cell populations. As already mentioned above, the myeloid populations seem to have been an artefact of
+
+<--- Page Split --->
+
+insufficient vertical integration. Retrospectively, the integration used at that time by means of the cell ranger pipeline was not a sufficient approach. The MNN integration used now offers a much better approach to data integration.  
+
+4. The experiments in slice cultures should be described in more detail in the main text. Here, they state that they performed "myeloid depletion" when in fact they performed microglia depletion (as stated as header in the methods section).  
+
+We have tried to discuss the experimental part, especially the limitations, in detail. Our new data and analyses better support the reported results. The chemical depletion of microglia used here, the myeloid population in the human brain slice, has already been described in detail our previous work<sup>1,2</sup>.  
+
+The effect size of myeloid/microglia depletion on IL10 production is rather modest. One missing control is depletion of other cell types within the slice culture and measurement of the effect on IL10 to exclude the possibility that there are other major sources of IL10 production (which is likely).  
+
+Indeed, we agree that there are other sources of IL10 release in the tumor microenvironment. Neurons, astrocytes, and oligodendrocytes are potential candidates for IL10 release, but protocols for depletion of these cell types have not yet been established. To address this question, we attempted to establish cell- specific depletion for astrocytes. Unfortunately, we failed with this because the toxicity of astrocyte depletion is too high. Our results with IL10 inhibition confirm the downstream mechanism but cannot conclusively resolve the question of which cell type should also be considered as an IL10 source. At least our data suggest that a large fraction of IL10 is derived from HMOX1- positive myeloid cells.  
+
+Furthermore, it is surprising that IL2 production, but not IFNG production increases during T cell depletion - the authors should offer potential explanations as this argues against reinvigoration of T cell poly- functionality.  
+
+We agree with the reviewer that the IFNG signal is difficult to explain. When we examined the raw signal, we found generally high levels (including the negative control) of IFNG and other cytokines in the ELISA, suggesting a potential technical problem. We fully reviewed the ELISA data and re- ran the assay. The new data provided a clearer picture. Also in the new data, we see only a small baseline effect on IL10 after depletion of microglia. We assume that within our slice model the reactive transformation of microglia is necessary to upregulate IL10 release. Therefore, differences in the "no- tumor" slices were not observed.  
+
+The results shown in 4j confirm that IL10 is an immunosuppressive cytokine, but do not substantiate claims that this is mediated by myeloid cells. Again, perplexing that no change in interferon gamma is seen.  
+
+As mentioned above, we are limited within our model to elucidate all potential sources of IL10 release. Our data confirms that a significant part originates from myeloid cells.
+
+<--- Page Split --->
+
+The single patient study is encouraging - was this pre-/post- comparison performed after single- agent therapy with ruxolitinib or was there a combination used? If the latter, it is possible that observed effects are due to other treatment constituents (see comment 1).  
+
+We used monotherapy in a neoadjuvant setting. However, the patient was pretreated with RT+TMZ +CCNU (CeTeG protocol) and received TTF (not in parallel to ruxolitinib). Resection of the tumor was subsequently performed after 6 weeks of ruxolitinib monotherapy. For analysis, we were able to perform staining for direct pre/post- treatment. Unfortunately, single cell sequencing could not be performed from the very small biopsy sample (before therapy), but only from the surgery sample (after treatment). We added a more detailed description and illustrations in the updated manuscript.
+
+<--- Page Split --->
+
+## Bibliography  
+
+1. Ravi, V. M. et al. Human organotypic brain slice culture: a novel framework for environmental research in neuro-oncology. Life Sci. Alliance 2, (2019).  
+2. Henrik Heiland, D. et al. Tumor-associated reactive astrocytes aid the evolution of immunosuppressive environment in glioblastoma. Nat. Commun. 10, 2541 (2019).  
+3. Hao, Y., Hao, S. & Andersen, E. Integrated analysis of multimodal single-cell data. Nissen  
+4. Mathewson, N. D. et al. Inhibitory CD161 receptor identified in glioma-infiltrating T cells by single-cell analysis. Cell 184, 1281–1298.e26 (2021).
+
+<--- Page Split --->
+
+## REVIEWER COMMENTS  
+
+Reviewer #1 (Remarks to the Author):  
+
+The authors have substantially improved the presentation of the hypotheses, analyses, and data in the revised manuscript, and have incorporated several new analyses that fill some of the gaps in the previous version. In my opinion, the revised manuscript is suited for publication.  
+
+I would only like to point some typos and small suggestions to improve the clarity of some parts:  
+
+- Line 123: "reference datasets" -> "reference dataset"  
+- Line 137: "Supplementary Figure 1b-c" -> "Figure 1b-c"  
+- Line 189: "Figure 2b" -> "Figure 2c"  
+- Line 223: "Figure 3d" -> "Figure 3f"  
+- Line 309: "Figure 5b" -> "Figure 5c"  
+- Line 389: "Benjamini-Hochberger" -> "Benjamini-Hochberg"  
+
+- Line 316: the notation used for the two clusters of myeloid cells (aM\Phi and bM\Phi) does not match with the notation used in Supplementary Fig. 5.  
+
+- In figure 1d,e it is unclear what the authors mean by "z-scored Gene Expression". z-scores are not bounded between 0 and 1, so I suspect the authors refer to something else or they have rescaled the z-scores in some way to lie between 0 and 1. I would suggest clarifying the normalization used in the figure legend and in the methods section.  
+
+- The specific gene sets from MSigDB v7 that were used should be specified in the methods section.  
+
+- I find the notation used in Fig. 5c and other figures (circle color + size) to be confusing. For example, in Fig. 5c the Mes-like correlation of #URK_S3 seems larger than the Mes-like correlation of #UKF_S2 based on the color of the circle, but smaller or equal based on the size of the circle.  
+
+- It would be helpful to add more details in the legend of Fig. 5b. For example, what does each of the colors in the figure denotes?  
+
+Reviewer #2 (Remarks to the Author):  
+
+In their revised manuscript titled "Crosstalk between lymphoid and myeloid cells orchestrates glioblastoma immunity through Interleukin 10 signaling", investigators coupled scRNAseq and stRNAseq to query the tumor microenvironment of 8 treatment naive glioblastoma patients. Ligand/receptor interactions were identified using a novel algorithm, nearest functionally connected neighbors (NFCN), ultimately identifying HMOX1+ myeloid cells as a major source of IL- 10. A T cell exhaustion phenotype was linked to the HMOX1+ myeloid cells and validated with both stRNAseq and an ex vivo neocortical system. In this revision, authors addressed concerns regarding the limitations of ex vivo experiments in validation of in silico findings. Additionally, authors increased the detail in many of the results sections to clarify relevant findings. These revisions greatly improve the quality and impact of the manuscript.
+
+<--- Page Split --->
+
+Reviewer #3 (Remarks to the Author):  
+
+The authors have done a very good job revising the manuscript and addressing my comments and suggestions.  
+
+I would strongly encourage them to highlight some of the technical and experimental challenges they had in their revision, as this might be of importance for future studies by their and other groups.  
+
+Benjamin Izar
+
+<--- Page Split --->
+
+## Point - by - Point Revision 2  
+
+D. H. Heiland  
+
+We thank all reviewers for their time and effort in evaluating our manuscript and appreciate the positive feedback on our project.  
+
+Reviewer #1 (Remarks to the Author):  
+
+The authors have substantially improved the presentation of the hypotheses, analyses, and data in the revised manuscript, and have incorporated several new analyses that fill some of the gaps in the previous version. In my opinion, the revised manuscript is suited for publication.  
+
+I would only like to point some typos and small suggestions to improve the clarity of some parts:  
+
+- Line 123: "reference datasets" -> "reference dataset"  
+
+Changed  
+
+- Line 137: "Supplementary Figure 1b-c" -> "Figure 1b-c"  
+
+Changed  
+
+- Line 189: "Figure 2b" -> "Figure 2c"  
+
+Changed  
+
+- Line 223: "Figure 3d" -> "Figure 3f"  
+
+Changed  
+
+- Line 309: "Figure 5b" -> "Figure 5c"  
+
+Changed  
+
+- Line 389: "Benjamini- Hochberger" -> "Benjamini-Hochberg"  
+
+Changed  
+
+- Line 316: the notation used for the two clusters of myeloid cells (aM\Phi and bM\Phi) does not match with the notation used in Supplementary Fig. 5.  
+
+Changed in the new supplementary file  
+
+- In figure 1d,e it is unclear what the authors mean by "z-scored Gene Expression". z-scores are not bounded between 0 and 1, so I suspect the authors refer to something else or they have rescaled the z-scores in some way to lie between 0 and 1. I would suggest clarifying the normalization used in the figure legend and in the methods section.  
+
+It is normalized gene expression, we changed the figure description  
+
+- The specific gene sets from MSigDB v7 that were used should be specified in the methods section.  
+
+The gene sets are implemented in the method part.  
+
+- I find the notation used in Fig. 5c and other figures (circle color + size) to be confusing. For example, in Fig. 5c the Mes-like correlation of #URK_S3 seems larger than the Mes-like correlation of #UKF_S2 based on the color of the circle, but smaller or equal based on the size of the circle.
+
+<--- Page Split --->
+
+It was based on the fact that plots were created individually, we changed this.  
+
+- It would be helpful to add more details in the legend of Fig. 5b. For example, what does each of the colors in the figure denotes?  
+
+We added a description in the figure.  
+
+Reviewer #2 (Remarks to the Author):  
+
+In their revised manuscript titled "Crosstalk between lymphoid and myeloid cells orchestrates glioblastoma immunity through Interleukin 10 signaling", investigators coupled scRNAseq and stRNAseq to query the tumor microenvironment of 8 treatment naive glioblastoma patients. Ligand/receptor interactions were identified using a novel algorithm, nearest functionally connected neighbors (NFCN), ultimately identifying HMOX1+ myeloid cells as a major source of IL- 10. A T cell exhaustion phenotype was linked to the HMOX1+ myeloid cells and validated with both stRNAseq and an ex vivo neocortical system. In this revision, authors addressed concerns regarding the limitations of ex vivo experiments in validation of in silico findings. Additionally, authors increased the detail in many of the results sections to clarify relevant findings. These revisions greatly improve the quality and impact of the manuscript.  
+
+Reviewer #3 (Remarks to the Author):  
+
+The authors have done a very good job revising the manuscript and addressing my comments and suggestions. I would strongly encourage them to highlight some of the technical and experimental challenges they had in their revision, as this might be of importance for future studies by their and other groups.  
+
+R2 & R3: Thank you for the positive response. We are pleased that the reviewers appreciated the considerable effort and the change during the revision process. In our opinion, a description of these changes in the discussion clearly hampers the red- thread and takes the focus away from the scientific results. In our opinion, the detailed process can be reconstructed very well from published reviewers' comments, so we have refrained from making a statement in the discussion part.
+
+<--- Page Split --->

peer_reviews/126b45d3c59304c8cdcaeff2fb15da92849086ba5d613f1f0d668d95576aa34f/supplementary_0_Peer Review File/supplementary_0_Peer Review File_det.mmd

@@ -0,0 +1,589 @@
+<|ref|>title<|/ref|><|det|>[[61, 40, 508, 90]]<|/det|>
+# nature portfolio  
+
+<|ref|>title<|/ref|><|det|>[[66, 110, 362, 139]]<|/det|>
+# Peer Review File  
+
+<|ref|>text<|/ref|><|det|>[[88, 154, 910, 210]]<|/det|>
+T- Cell Dysfunction in the Glioblastoma Microenvironment is Mediated by Myeloid Cells Releasing Interleukin- 10  
+
+<|ref|>image<|/ref|><|det|>[[56, 732, 240, 781]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[250, 732, 912, 784]]<|/det|>
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[119, 83, 312, 98]]<|/det|>
+## REVIEWER COMMENTS  
+
+<|ref|>text<|/ref|><|det|>[[120, 112, 402, 128]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[118, 142, 879, 262]]<|/det|>
+This manuscript studies the crosstalk between infiltrating microglia/macrophages and T- cells in the immunosuppressive microenvironment of glioblastoma. This is an important topic that can offer further insights into the mechanisms of resistance to anti- PD1 immunotherapy (Zhao et al. Nat. Medicine 25 (2019)). This work is a continuation of previously published work (Herik Heiland et al., Nat. Commun. 10 (2019)), where the authors found that the crosstalk between microglia cells and reactive astrocytes is responsible for upregulating IL- 10 release in glioblastoma through JAK/STAT signaling. In this manuscript, the authors show that IL- 10 secreted by HMOX1+ myeloid cells is responsible for inducing a dysfunctional state in T cells infiltrating mesenchymal regions of the tumor.  
+
+<|ref|>text<|/ref|><|det|>[[120, 275, 759, 290]]<|/det|>
+There are several aspects that in my opinion need to be addressed before publication:  
+
+<|ref|>text<|/ref|><|det|>[[119, 304, 868, 364]]<|/det|>
+1. Some of the statements in the Abstract, Introduction, and Discussion do not seem to have clear support in the data and analyses that are presented in the manuscript. Either those statements need to be more carefully crafted or the data and analyses supporting them need to be more clearly presented:  
+
+<|ref|>text<|/ref|><|det|>[[119, 377, 872, 437]]<|/det|>
+1a. Lines 375-378: the authors state that their analysis indicates that the dysfunctional state of T cells appears to be a transient state. However, this reviewer was unable to find any data or analyses supporting the transitory character of dysfunctional T cell states in subsection "Dysfunctional State of T cells is Driven by IL-10 Signaling".  
+
+<|ref|>text<|/ref|><|det|>[[118, 450, 873, 570]]<|/det|>
+1b. Lines 58, 117-119, and 394-396: the authors mention that HMOX1+ myeloid cells co-localize spatially with the mesenchymal signature of glioblastoma. However, Figs. 3f,g only show the spotwise correlation between dysfunctional T cell gene expression markers (HAVCR2 and LAG3) and the mesenchymal gene expression signature, but as far as I see they do not present any analysis of the spot-wise correlation between HMOX1+ myeloid cell markers and the mesenchymal gene expression signature. Similarly, in subsection "T cell Activation and Exhaustion Reveals Spatial Heterogeneity and Association with Glioblastoma Subtypes" there seem to be no results about the spatial pattern of HMOX1+ myeloid cells.  
+
+<|ref|>text<|/ref|><|det|>[[118, 583, 864, 644]]<|/det|>
+1c. Lines 403-405. The authors state that their experiments with the ex vivo neocortical GBM model confirm that HMOX1+ myeloid cells cause a reduction of effector T cells. However, in these experiments both HMOX1+ and HMOX1- myeloid cells are depleted using clodronate. In absence of other data, these experiments only show that myeloid cells cause a reduction of effector T cells.  
+
+<|ref|>text<|/ref|><|det|>[[118, 656, 878, 717]]<|/det|>
+1d. Lines 352-356 and 414-418, and Fig 4q-r. It is unclear what comparison was performed here. How did the authors determine that there is a significant enrichment for activated T cells and B cells upon JAK/STAT inhibition in the patient? A more detailed explanation of the comparison and the assumptions that were made would be useful here.  
+
+<|ref|>text<|/ref|><|det|>[[117, 730, 875, 880]]<|/det|>
+2. The authors introduce the "nearest functionally connected neighbor" algorithm to infer candidate paracrine interactions from the single-cell RNA-seq data. Since the performance of this algorithm has not been rigorously evaluated, it is hard to know how reliable the results of this method are in general. In Supplementary Fig. 5, the authors show that the more conventional and established algorithm NicheNet (Browaeys et al. Nat. Methods 17 (2020)) also finds the candidate interaction between myeloid cells (expressing IL10) and T-cells (expressing IL10RA). However, NicheNet does not show that this interaction is specific to HOMX1+ myeloid cells. Can the authors present any other evidence from the single-cell RNA-seq data to support their statement that HOMX1+ cells are mostly responsible for this interaction (e.g. correlation between HOMX1+ and IL10 expression within myeloid cells)? It would be also useful to show the UMAP representation labeled by HMOX1 expression.  
+
+<|ref|>text<|/ref|><|det|>[[115, 893, 860, 909]]<|/det|>
+3. The description of the "nearest functionally connected neighbor" algorithm in the Methods section
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[118, 82, 848, 112]]<|/det|>
+lacks much technical detail. It would be useful to include details about the models, fitting methods, etc., and rewrite the description of the algorithm more carefully.  
+
+<|ref|>text<|/ref|><|det|>[[118, 127, 868, 262]]<|/det|>
+4. I found the main figures to be unnecessarily complex with 10-18 panels each. The authors might consider keeping the panels that convey the main results and moving the rest of the panels to supplementary figures. There are also several typos that need to be corrected. For example, the panels in Fig. 3 are mismatched with the figure legend (e.g. 3c and 3d seem to be exchanged) and with the main text (lines 283-310). Supplementary Fig. 6 includes a caption "Supplementary Fig. 5" which should be removed. The legend of Supplementary Fig. 3 says "Dimensional reduction (UMAP) of gene expression of the different simulation experiments". However, what is shown in the figure seems to be the UMAP of the single-cell RNA-seq data from the patients colored by the imputed stimulation signatures.  
+
+<|ref|>text<|/ref|><|det|>[[120, 305, 401, 320]]<|/det|>
+Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[118, 334, 880, 380]]<|/det|>
+NCOMMS- 21- 07876- T Crosstalk between lymphoid and myeloid cells orchestrates glioblastoma immunity through Interleukin 10 signaling  
+
+<|ref|>text<|/ref|><|det|>[[118, 393, 878, 585]]<|/det|>
+In their manuscript titled "Crosstalk between lymphoid and myeloid cells orchestrates glioblastoma immunity through Interleukin 10 signaling", investigators leveraged single- cell and spatial transcriptomics to infer cellular crosstalk between macrophages/microglia and T cells within human GBM samples. From eight GBM samples, 21 clusters were identified, where many of these clusters represented macrophages and microglia. Sub- clustering of the T cell cluster revealed different activation states that were then mapped with pseudo- time and RNA velocity analyses. This provided a differentiation map from naive to terminally exhausted, with an intermediate highly proliferative state. Additionally, by combining these techniques, authors identified an association between the mesenchymal GBM subtype and T cell exhaustion. Furthermore, authors developed a new model, nearest functional connected neighbor, to identify ligand/receptor interactions from scRNAseq data. Computational analyses are thorough and novel; however, the follow- up mechanistic studies, using an in vitro neocortical model, were limited in their support of computational findings. Primary concerns are related to the impact of the mechanistic studies. These concerns are detailed below:  
+
+<|ref|>text<|/ref|><|det|>[[118, 599, 870, 717]]<|/det|>
+1. A major limitation is the in vitro neocortical model used in Figure 4. Authors state that this model replicates the tumor microenvironment because it is derived from brain tissue; however, during tumorigenesis the TME is largely shaped by the infiltrating immune cells (1), which are absent in this model. Authors should comment on the "myeloid" cells that exist within the neocortical tissue that are being evaluated in figure 4, which should primarily be microglia and not the HMOX1+ macrophages from the computational studies. Along these lines, a quantification of tumor growth in the myeloid depleted condition is important, as myeloid cells generally promote tumor growth, so their absence alone may have an effect on tumor growth independent of T cells or IL10 inhibition.  
+
+<|ref|>text<|/ref|><|det|>[[118, 731, 875, 880]]<|/det|>
+2. Additionally, the short incubation time (3 days) from tumor injection and subsequent T cell transfer may not provide sufficient time for meaningful interactions and subsequent functional outputs to occur, for example T cell exhaustion. Regarding the transferred T cells, it was not clearly stated whether there is selection for tumor-specific T cells from patient blood, therefore T cells that are injected into the neocortical model may or may not react to the tumor. Evidence of T cell recognition of tumor cells through killing or activation is necessary to increase the impact of this model. Much of T cell exhaustion biology is ignored in this model, such as the conditions and locations under which priming occur and trafficking to the tumor site. Therefore, conclusions that can be drawn from this model are limited. Furthermore, authors should address potential allogeneic reactivity to the cell line BTSC#233 by patient T cells.  
+
+<|ref|>text<|/ref|><|det|>[[118, 895, 808, 910]]<|/det|>
+3. Although important for spatial information, the immunofluorescence images shown without
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[118, 83, 868, 158]]<|/det|>
+quantifications in figure 4 are not sufficient to validate the computational studies. For example, in figure 4g, TIM3 is used to identify "exhausted T cells", but it has been shown that TIM3 can also be a marker of terminal effector differentiation. Therefore, this would be more convincing if other parameters were used to identify this population, such as PD1 or other functional studies showing T cell activity.  
+
+<|ref|>text<|/ref|><|det|>[[118, 171, 875, 247]]<|/det|>
+4. When looking at the contribution of different patients to the final clusters, it is apparent that many clusters are specific to single patients. In particular, \(80\%\) of the HMOX1+ group is made up of a single patient. What impact does this have on the broader applicability of these findings? Would this bias for a single patient carry over into the downstream analyses? Is this to be expected whenever pooling human samples?  
+
+<|ref|>text<|/ref|><|det|>[[118, 260, 800, 290]]<|/det|>
+5. The approval for the use of an un-licensed drug in a GBM patient needs to be specifically addressed within the ethics section.  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 305, 258, 318]]<|/det|>
+## 6. Minor concerns:  
+
+<|ref|>text<|/ref|><|det|>[[118, 319, 860, 420]]<|/det|>
+a. The conclusion that myeloid and lymphoid interactions lead to T cell dysfunction through IL10 is vague and not unoriginal. Secretion of IL10 by myeloid cells is not a novel finding, nor is the role of IL10 in T cell dysfunction(2,3).  
+b. Using in vitro stimulated T cells to compare with in vivo T cells coming from a tumor ignores much of the complexity in signals that is occurring intratumorally.  
+c. Using a second cell line in the in vitro neocortical studies would increase the impact of related findings.  
+
+<|ref|>text<|/ref|><|det|>[[118, 435, 872, 555]]<|/det|>
+In general, authors show novel and interesting computational analyses from cutting edge techniques, however they lack substance in their follow up mechanistic studies. The use of pseudotime and RNA velocity to interrogate T cell activation states and pair them with GBM subtypes and spatial information are intriguing. Additionally, the nearest functionally connected neighbor algorithm will add to the expanding pool of resources for inferring intercellular communication from scRNAseq data. Follow-up studies are limited in their support of computational findings (IL- 10 mediated T cell exhaustion); therefore, additional validation studies are required, or the focus of the story should shift to highlight the novel bioinformatic analyses.  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 585, 202, 598]]<|/det|>
+## Reference:  
+
+<|ref|>text<|/ref|><|det|>[[117, 599, 870, 689]]<|/det|>
+1. Salmon, H., Remark, R., Gnjatic, S. and Merad, M., 2019. Host tissue determinants of tumour immunity. Nature Reviews Cancer, 19(4), pp.215-227.  
+2. McLane, L.M., Abdel-Hakeem, M.S. and Wherry, E.J., 2019. CD8 T cell exhaustion during chronic viral infection and cancer. Annual review of immunology, 37, pp.457-495.  
+3. Quail, D.F. and Joyce, J.A., 2017. The microenvironmental landscape of brain tumors. Cancer cell, 31(3), pp.326-341.  
+
+<|ref|>text<|/ref|><|det|>[[119, 747, 402, 761]]<|/det|>
+Reviewer #3 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[118, 775, 870, 910]]<|/det|>
+Ravi, Neidert, Will et al present a single- cell RNA- sequencing study of tumor- infiltrating lymphocytes of patients with GBM. The profile 8 patients and additionally show data for 3 additional patients using spatial profiling, all using 10x platforms. Using established and novel analytical tools to infer trajectories of cell differentiation, they identify variability among T cells; among other, they find subclusters of CD8+ T cells with high expression of dysfunction marker TIM3 (HAVCR2) and cells with a hypoxia signature with distinct trajectories. They correlate these signatures with signatures of T cells collected following in vitro stimulation with different cytokines, including IL2, IFNG and IL10, arguing that IL10 stimulated cells have a lower activation score compared to other cells identified as effector cells. Using spatial RNA- seq of three tumors, they find an association of tumor- mesenchymal niches
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[118, 82, 875, 275]]<|/det|>
+and infiltration of exhausted T cells (TIM3/LAG3), suggesting that niches of dysfunctional T cells may be in part explained by cancer cell intrinsic features. In order to understand the origin of IL10 and to solidify the role of T cells as IL10 recipients, they present a analytical framework that infers ligandreceptor interactions using several constraints, and demonstrate that myeloid cells (CD163+, HMOX1+) are a main source of IL10. To begin validating this finding, they use slice cultures that they deplete of myeloid cells and show that depletion of myeloid cells results in reduction of IL10 (in the presence of tumor cells); in these models, they then co- culture autologous T cells and show that depletion of myeloid cells in slice cultures results in increased IL2, but not IFNG protein abundance. Incubation of T cells with an IL10R- inhibitor prior to co- culture with tissues results in increased IL2 production in T cells. Because the JAK/STAT pathway is downstream of IL10 signaling, they use ruxolitinib (a selective JAK1/2 inhibitor) first i slice model showing increased IL2, and then use this drug in the neo- adjuvant therapy of a patient with GBM, followed by analysis of the surgical specimen, which shows activation of T cells.  
+
+<|ref|>text<|/ref|><|det|>[[119, 289, 860, 349]]<|/det|>
+GBM is a disease with extremely poor prognosis, and therapeutic development has in part been hampered by limited understanding of the tumor microenvironment; as such, the study of potentially high importance. However, several aspects raised significant concerns and reduced enthusiasm for this study, and need to be addressed,  
+
+<|ref|>text<|/ref|><|det|>[[119, 364, 216, 378]]<|/det|>
+Major points:  
+
+<|ref|>text<|/ref|><|det|>[[118, 393, 876, 525]]<|/det|>
+1. Nowhere in the manuscript do the authors describe the characteristics of the patient tumors used for either single-cell sequencing of spatial sequencing. Are these all treatment naive tumors? where they exposed to different therapies (radiation, chemotherapy, immunotherapy, investigational drugs) - this will have a dramatic impact on the measured T cell phenotypes and in and of itself could describe variability seen in the data set. The authors should describe basic demographics and treatment history; it is not reasonable to request from the authors to attempt to account for variability based on basic demographics (this, and virtually any single-cell study would be underpowered), but they should show major analyses/findings in the context of different therapies received to exclude the possibility noted above.  
+
+<|ref|>text<|/ref|><|det|>[[118, 540, 876, 701]]<|/det|>
+2. Technical quality: it is somewhat surprising that the authors only recover \(\sim 1000\) unique genes per cell with only \(\sim 2300\) unique molecular identifiers - this is not on par with the quality described and raises concerns regarding data quality; this is particularly surprising as they used the 3.1 chemistry which performs better than prior chemistries; in fact recent studies performing profiling from frozen tissues even achieved similar or better quality compared to this present study (Slyper, Nature Medicine, 2020). Furthermore, some of the clusters described might be artifactual due to tissue processing (e.g. "hypoxia cluster")- this possibility should be addressed using available data sets systematically investigating such artifacts (e.g. Ido Amit laboratory). Furthermore, the authors should comment on the technical quality. An additional embedding showing the UML count for major analyses should be shown in the supplement to exclude the possibility of technical artifacts as drivers of clustering.  
+
+<|ref|>text<|/ref|><|det|>[[118, 716, 878, 820]]<|/det|>
+3. There is no statistical evaluation of the inference made in Figure 3f - the authors should provide this; in this same figure, they also show that CCL2 myeloid cells are scoring highly, which is a gene considered to be an immunostimulatory gene/protein - how do they reconcile this? This brings up the question about a more nuanced annotation of the myeloid cells beyond monocytes, macrophages and microglia. The effect size of the gene set enrichment analysis in 3h is very underwhelming. In fact, throughout this section and the studies shown in Figure 4, the effect sizes are very small with in part borderline significance, raising the question of biological significance of these findings.  
+
+<|ref|>text<|/ref|><|det|>[[118, 834, 878, 909]]<|/det|>
+4. The experiments in slice cultures should be described in more detail in the main text. Here, they state that they performed "myeloid depletion" when in fact they performed microglia depletion (as stated as header in the methods section). The effect size of myeloid/microglia depletion on IL10 production is rather modest. One missing control is depletion of other cell types within the slice culture and measurement of the effect on IL10 to exclude the possibility that there are other major sources of
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[118, 83, 860, 201]]<|/det|>
+IL10 production (which is likely). Furthermore, it is surprising that IL2 production, but not IFNG production increases during T cell depletion - the authors should offer potential explanations as this argues against reinvigoration of T cell poly- functionality. The results shown in 4j confirm that IL10 is an immunosuppressive cytokine, but do not substantiate claims that this is medicated by myeloid cells. Again, perplexing that no change in interferon gamma is seen. The single patient study is encouraging - was this pre- /post- comparison performed after single- agent therapy with ruxolitinib or was there a combination used? If the latter, it is possible that observed effects are due to other treatment constituents (see comment 1).
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[120, 86, 395, 112]]<|/det|>
+## Point - by - Point  
+
+<|ref|>text<|/ref|><|det|>[[120, 124, 222, 138]]<|/det|>
+D. H. Heiland  
+
+<|ref|>text<|/ref|><|det|>[[118, 164, 881, 243]]<|/det|>
+We thank all reviewers for their time and effort in evaluating our manuscript and appreciate the positive feedback on our project. We have tried to mitigate the issues highlighted by the reviewers, which has led to a significant improvement in the quality of our manuscript. The following main changes have been made in this context:  
+
+<|ref|>sub_title<|/ref|><|det|>[[120, 268, 448, 283]]<|/det|>
+## 1. Quality of the scRNA-seq experiments.  
+
+<|ref|>text<|/ref|><|det|>[[118, 288, 880, 325]]<|/det|>
+By resequencing the libraries, the quality of the entire dataset was significantly improved and the number of detected genes as well as the UMIs per cell were significantly increased.  
+
+<|ref|>sub_title<|/ref|><|det|>[[119, 350, 374, 365]]<|/det|>
+## 2. Analysis and data integration  
+
+<|ref|>text<|/ref|><|det|>[[118, 370, 880, 427]]<|/det|>
+By applying more advanced algorithms for horizonal and vertical data integration as well as cell type alignment, we were able to present a clearer picture of T cell diversity in the tumors. We separated CD4 and CD8 positive T cells for all downstream analysis.  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 452, 456, 467]]<|/det|>
+## 3. Avoid overfitting in the NFCN algorithm  
+
+<|ref|>text<|/ref|><|det|>[[118, 472, 881, 550]]<|/det|>
+Our cell communication algorithm has been optimized to reduce potential overfitting and improve prediction. For this purpose, we integrated multiple prediction/validation layers and external algorithms. The new version is also compatible with the conventional scRNA- seq tools (Seurat) and available as an R package (NFCN2).  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 575, 500, 590]]<|/det|>
+## 4. Structure of the manuscript and presentation  
+
+<|ref|>text<|/ref|><|det|>[[118, 595, 880, 632]]<|/det|>
+We restructured our manuscript to present clear hypothesis- driven argumentation and pointed out limitations and ambiguities. The illustrations have been simplified to improve general understanding.  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 656, 274, 670]]<|/det|>
+## 5. Validation model  
+
+<|ref|>text<|/ref|><|det|>[[118, 676, 880, 713]]<|/det|>
+Our experimental model is not without limitations, which are discussed in detail. We performed new experiments and analysis to improve the experimental validation.  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 738, 265, 752]]<|/det|>
+## 6. Clinical Dataset  
+
+<|ref|>text<|/ref|><|det|>[[118, 758, 880, 795]]<|/det|>
+Our in- vivo dataset is now described in detail, and we were able to generate further data to strengthen our hypothesis.  
+
+<|ref|>text<|/ref|><|det|>[[115, 820, 797, 836]]<|/det|>
+In order to discuss the reviewer comments in detail, we provide a point- by- point discussion.
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[119, 82, 420, 98]]<|/det|>
+## Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[117, 123, 882, 284]]<|/det|>
+This manuscript studies the crosstalk between infiltrating microglia/macrophages and T- cells in the immunosuppressive microenvironment of glioblastoma. This is an important topic that can offer further insights into the mechanisms of resistance to anti- PD1 immunotherapy (Zhao et al. Nat. Medicine 25 (2019)). This work is a continuation of previously published work (Henrik Heiland et al., Nat. Commun. 10 (2019)), where the authors found that the crosstalk between microglia cells and reactive astrocytes is responsible for upregulating IL- 10 release in glioblastoma through JAK/STAT signaling. In this manuscript, the authors show that IL- 10 secreted by HMOX1+ myeloid cells is responsible for inducing a dysfunctional state in T cells infiltrating mesenchymal regions of the tumor.  
+
+<|ref|>text<|/ref|><|det|>[[118, 308, 880, 345]]<|/det|>
+We would like to thank the reviewer for his time and comments leading to an improvement of the manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[119, 370, 760, 386]]<|/det|>
+There are several aspects that in my opinion need to be addressed before publication:  
+
+<|ref|>text<|/ref|><|det|>[[118, 410, 881, 469]]<|/det|>
+1. Some of the statements in the Abstract, Introduction, and Discussion do not seem to have clear support in the data and analyses that are presented in the manuscript. Either those statements need to be more carefully crafted or the data and analyses supporting them need to be more clearly presented:  
+
+<|ref|>text<|/ref|><|det|>[[118, 492, 881, 591]]<|/det|>
+We have substantially revised our argumentation to better support the data presented and to clearly define our hypotheses. By improving the scRNA-seq datasets and analysis approaches, some of our previously stated hypotheses have been relativized. The major difference compared to our previous manuscript is a separation of CD8 and CD4 positive T cells for all further sub analysis. Our new data incorporated novel aspects of the underlying mechanism of tumor-associated T cell response.  
+
+<|ref|>text<|/ref|><|det|>[[118, 615, 881, 692]]<|/det|>
+1a. Lines 375- 378: the authors state that their analysis indicates that the dysfunctional state of T cells appears to be a transient state. However, this reviewer was unable to find any data or analyses supporting the transitory character of dysfunctional T cell states in subsection "Dysfunctional State of T cells is Driven by IL- 10 Signaling".  
+
+<|ref|>text<|/ref|><|det|>[[118, 697, 881, 795]]<|/det|>
+In our revised version, we performed model integration of RNA- velocity and lineage tree reconstruction to improve the exploration of state specific pathway activation. We found that IL10 response was highly correlated with the expression of exhaustion programs in two T cell clusters. We rewrote this part of the manuscript to improve understanding and removed statements that are no longer supported or have caused confusion.  
+
+<|ref|>text<|/ref|><|det|>[[118, 820, 881, 898]]<|/det|>
+1b. Lines 58, 117- 119, and 394- 396: the authors mention that HMOX1+ myeloid cells co- localize spatially with the mesenchymal signature of glioblastoma. However, Figs. 3f,g only show the spot- wise correlation between dysfunctional T cell gene expression markers (HAVCR2 and LAG3) and the mesenchymal gene expression signature, but as far as I see they do not present any analysis of the
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[118, 82, 881, 161]]<|/det|>
+spot- wise correlation between HMOX1+ myeloid cell markers and the mesenchymal gene expression signature. Similarly, in subsection "T cell Activation and Exhaustion Reveals Spatial Heterogeneity and Association with Glioblastoma Subtypes" there seem to be no results about the spatial pattern of HMOX1+ myeloid cells.  
+
+<|ref|>text<|/ref|><|det|>[[118, 185, 881, 243]]<|/det|>
+We thank the reviewer for picking up on this lack of clarity in presentation. We have improved the text and figure legends for clarity. In our revised version of the manuscript, we added spot- wise correlations to support our hypothesis as well as spatial data analysis of the model.  
+
+<|ref|>text<|/ref|><|det|>[[118, 266, 881, 345]]<|/det|>
+1c. Lines 403- 405. The authors state that their experiments with the ex vivo neocortical GBM model confirm that HMOX1+ myeloid cells cause a reduction of effector T cells. However, in these experiments both HMOX1+ and HMOX1- myeloid cells are depleted using clotronate. In absence of other data, these experiments only show that myeloid cells cause a reduction of effector T cells.  
+
+<|ref|>text<|/ref|><|det|>[[117, 368, 882, 530]]<|/det|>
+Thank you for this helpful comment. It is indeed the case that we remove all myeloid cells and therefore are unable to differentiate HMOX1 pos/neg myeloid cells individually. Using our human model, we are currently not able to specifically target HMOX1 positive cells. We have described these limitations. To approach this limitation, we quantified the spatial distance of HMOX1- positive and - negative cells in our model and concluded that HMOX- positive cells are mainly localized in the proximity of the tumor. Thus, we assumed that HMOX1- negative cells were only present to a small extent within the tumor. However, in our more detailed analysis of the slice model, we discuss limitations and cofounders more detailed.  
+
+<|ref|>text<|/ref|><|det|>[[118, 553, 881, 632]]<|/det|>
+1d. Lines 352- 356 and 414- 418, and Fig 4q- r. It is unclear what comparison was performed here. How did the authors determine that there is a significant enrichment for activated T cells and B cells upon JAK/STAT inhibition in the patient? A more detailed explanation of the comparison and the assumptions that were made would be useful here.  
+
+<|ref|>text<|/ref|><|det|>[[118, 656, 880, 693]]<|/det|>
+This part was fully rewritten for an improved presentation of our hypothesis. The data are re- analyzed in accordance with our new findings.  
+
+<|ref|>text<|/ref|><|det|>[[117, 717, 882, 878]]<|/det|>
+2. The authors introduce the "nearest functionally connected neighbor" algorithm to infer candidate paracrine interactions from the single-cell RNA-seq data. Since the performance of this algorithm has not been rigorously evaluated, it is hard to know how reliable the results of this method are in general. In Supplementary Fig. 5, the authors show that the more conventional and established algorithm NicheNet (Browaeys et al. Nat. Methods 17 (2020)) also finds the candidate interaction between myeloid cells (expressing IL10) and T-cells (expressing IL10RA). However, NicheNet does not show that this interaction is specific to HOMX1+ myeloid cells. Can the authors present any other evidence from the single-cell RNA-seq data to support their statement that HOMX1+ cells are mostly responsible
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[118, 82, 880, 120]]<|/det|>
+for this interaction (e.g. correlation between HOMX1+ and IL10 expression within myeloid cells)? It would be also useful to show the UMAP representation labeled by HMOX1 expression.  
+
+<|ref|>text<|/ref|><|det|>[[117, 144, 882, 265]]<|/det|>
+In the new version of our "nearest functionally connected neighbor" (NFCN) algorithm, we implemented various new functions. In general, NFCN is built to quantify cellular interactions of a defined pathway (in our case the IL10- IL10R interaction). In comparison to NicheNet and CellChat, we inferred cellular connections based on the likelihood of cell pairs from the scRNA- seq dataset. Indeed, this quantification leads to overfitting as long as the ground truth is unknown. In order to overcome this problem, we redesigned the algorithm to integrate 3 data layers for improved prediction of cellular interactions.  
+
+<|ref|>text<|/ref|><|det|>[[118, 268, 603, 284]]<|/det|>
+1. Prediction of the cell-pair likelihood based on scRNA-seq data.  
+
+<|ref|>text<|/ref|><|det|>[[118, 289, 510, 304]]<|/det|>
+2. Deconvolution of Cell-Cell signaling from doublets  
+
+<|ref|>text<|/ref|><|det|>[[118, 309, 826, 325]]<|/det|>
+3. Integration of spatial resolved transcriptomics to confirm spatial juxta positioning of cell pairs.  
+
+<|ref|>text<|/ref|><|det|>[[117, 329, 881, 406]]<|/det|>
+We further integrated an unsupervised model using CellChat to infer the most common Cell- Cell interaction across clusters. Through our optimization, we tailored the model to predict cellular communication and reduced bias. We have added a supplementary result part to explain this model in detail.  
+
+<|ref|>text<|/ref|><|det|>[[118, 431, 880, 489]]<|/det|>
+3. The description of the "nearest functionally connected neighbor" algorithm in the Methods section lacks much technical detail. It would be useful to include details about the models, fitting methods, etc., and rewrite the description of the algorithm more carefully.  
+
+<|ref|>text<|/ref|><|det|>[[118, 513, 818, 529]]<|/det|>
+As mentioned in the answer above, we added supplementary results with detailed information.  
+
+<|ref|>text<|/ref|><|det|>[[117, 553, 882, 733]]<|/det|>
+4. I found the main figures to be unnecessarily complex with 10-18 panels each. The authors might consider keeping the panels that convey the main results and moving the rest of the panels to supplementary figures. There are also several typos that need to be corrected. For example, the panels in Fig. 3 are mismatched with the figure legend (e.g. 3c and 3d seem to be exchanged) and with the main text (lines 283-310). Supplementary Fig. 6 includes a caption "Supplementary Fig. 5" which should be removed. The legend of Supplementary Fig. 3 says "Dimensional reduction (UMAP) of gene expression of the different simulation experiments". However, what is shown in the figure seems to be the UMAP of the single-cell RNA-seq data from the patients colored by the imputed stimulation signatures.  
+
+<|ref|>text<|/ref|><|det|>[[118, 758, 880, 795]]<|/det|>
+Thanks for pointing out the typos and complexity of the figures. We have adapted the figures to facilitate ease of understanding.
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[119, 83, 420, 98]]<|/det|>
+## Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[117, 123, 882, 407]]<|/det|>
+Crosstalk between lymphoid and myeloid cells orchestrates glioblastoma immunity through Interleukin 10 signaling. In their manuscript titled "Crosstalk between lymphoid and myeloid cells orchestrates glioblastoma immunity through Interleukin 10 signaling", investigators leveraged single- cell and spatial transcriptomics to infer cellular crosstalk between macrophages/microglia and T cells within human GBM samples. From eight GBM samples, 21 clusters were identified, where many of these clusters represented macrophages and microglia. Sub- clustering of the T cell cluster revealed different activation states that were then mapped with pseudo- time and RNA velocity analyses. This provided a differentiation map from naive to terminally exhausted, with an intermediate highly proliferative state. Additionally, by combining these techniques, authors identified an association between the mesenchymal GBM subtype and T cell exhaustion. Furthermore, authors developed a new model, nearest functional connected neighbor, to identify ligand/receptor interactions from scRNAseq data. Computational analyses are thorough and novel; however, the follow- up mechanistic studies, using an in vitro neocortical model, were limited in their support of computational findings. Primary concerns are related to the impact of the mechanistic studies. These concerns are detailed below:  
+
+<|ref|>text<|/ref|><|det|>[[118, 430, 880, 468]]<|/det|>
+We would like to thank the reviewer for his time and comments leading to an improvement of the manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[117, 492, 882, 653]]<|/det|>
+1. A major limitation is the in vitro neocortical model used in Figure 4. Authors state that this model replicates the tumor microenvironment because it is derived from brain tissue; however, during tumorigenesis the TME is largely shaped by the infiltrating immune cells (1), which are absent in this model. Authors should comment on the "myeloid" cells that exist within the neocortical tissue that are being evaluated in figure 4, which should primarily be microglia and not the HMOX1+ macrophages from the computational studies. Along these lines, a quantification of tumor growth in the myeloid depleted condition is important, as myeloid cells generally promote tumor growth, so their absence alone may have an effect on tumor growth independent of T cells or IL10 inhibition.  
+
+<|ref|>text<|/ref|><|det|>[[117, 676, 882, 816]]<|/det|>
+Thank you for this valuable comment. We have addressed and discussed this limitation in detail. There is no doubt that the myeloid cells within the presented model are predominantly composed of microglial cells. However, these cells can also transform reactively and consequently become HMOX1 positive. HMOX1 positive microglial cells also play a crucial role in other pathologies such as traumatic brain injury and subarachnoid hemorrhage. Therefore, it would be safe to assume that although the model is limited, the specific role associated with HMOX1 expression can be associated with activated microglial cells. We have discussed this limitation in detail.  
+
+<|ref|>text<|/ref|><|det|>[[118, 840, 880, 899]]<|/det|>
+Regarding the quantification of tumor growth in myeloid depletion condition: This question is of high interest and our laboratory is currently working on this interaction. At the moment, we feel that the addition of this data will result in a loss of focus of the results presented in this manuscript.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[117, 82, 882, 284]]<|/det|>
+2. Additionally, the short incubation time (3 days) from tumor injection and subsequent T cell transfer may not provide sufficient time for meaningful interactions and subsequent functional outputs to occur, for example T cell exhaustion. Regarding the transferred T cells, it was not clearly stated whether there is selection for tumor-specific T cells from patient blood, therefore T cells that are injected into the neocortical model may or may not react to the tumor. Evidence of T cell recognition of tumor cells through killing or activation is necessary to increase the impact of this model. Much of T cell exhaustion biology is ignored in this model, such as the conditions and locations under which priming occur and trafficking to the tumor site. Therefore, conclusions that can be drawn from this model are limited. Furthermore, authors should address potential allogeneic reactivity to the cell line BTSC#233 by patient T cells.  
+
+<|ref|>text<|/ref|><|det|>[[118, 308, 880, 365]]<|/det|>
+Thank you for the detailed review of the model, which aids illustrating the various aspects, functionalities and limitations. Indeed, we are limited in the interpretation of our results. However, the following points deserve to be considered:  
+
+<|ref|>text<|/ref|><|det|>[[117, 389, 882, 530]]<|/det|>
+1. Regarding the first part of the question: We did not isolate tumor-specific T cells (mutation-associated neoantigens (MANA) associated TILs) from blood. In the context of brain tumors, to purify MANA-TILs is extremely challenging and only insufficiently possible using current methods. The aim of our model was to generate a T cell response and investigate the role of the tumor-associated microenvironment. Injection of a primary cell line which causes an allogeneic response is part of the model. Without this stimulus, a T cell response, as you mentioned above, is limited. This allogeneic reactivity should therefore be considered as intentional.  
+
+<|ref|>text<|/ref|><|det|>[[118, 554, 881, 632]]<|/det|>
+2) Regarding the second part of the question: Our data show that T cell activity (GZMB) in the tumor region becomes detectable after 3 days. (See data presented). The temporal dimensions of our slice model span a few days because the tumor infiltrates a large portion of the slice within 7 days. We have already reported tumor growth times in our previous publications<sup>1,2</sup>.  
+
+<|ref|>text<|/ref|><|det|>[[117, 656, 881, 755]]<|/det|>
+3. Although important for spatial information, the immunofluorescence images shown without quantifications in figure 4 are not sufficient to validate the computational studies. For example, in figure 4g, TIM3 is used to identify "exhausted T cells", but it has been shown that TIM3 can also be a marker of terminal effector differentiation. Therefore, this would be more convincing if other parameters were used to identify this population, such as PD1 or other functional studies showing T cell activity.  
+
+<|ref|>text<|/ref|><|det|>[[118, 779, 880, 836]]<|/det|>
+We added a more sophisticated validation of the imaging results. We choose Tim3 to confirm the results from the computational studies, in which the tissue resident memory cluster revealed the strongest enrichment of exhaustion markers.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[117, 102, 881, 202]]<|/det|>
+4. When looking at the contribution of different patients to the final clusters, it is apparent that many clusters are specific to single patients. In particular, \(80\%\) of the HMOX1+ group is made up of a single patient. What impact does this have on the broader applicability of these findings? Would this bias for a single patient carry over into the downstream analyses? Is this to be expected whenever pooling human samples?  
+
+<|ref|>text<|/ref|><|det|>[[118, 225, 787, 243]]<|/det|>
+This problem was based on the vertical integration algorithm which has been fully revised.  
+
+<|ref|>text<|/ref|><|det|>[[118, 266, 880, 305]]<|/det|>
+5. The approval for the use of an un-licensed drug in a GBM patient needs to be specifically addressed within the ethics section.  
+
+<|ref|>text<|/ref|><|det|>[[118, 328, 880, 366]]<|/det|>
+The treatment was performed as part of the "Compassionate Use" program (RL 2001/83/EG VO 726/2004). We added explanations in the manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[118, 389, 880, 448]]<|/det|>
+6. Minor concerns: a. The conclusion that myeloid and lymphoid interactions lead to T cell dysfunction through IL10 is vague and not unoriginal. Secretion of IL10 by myeloid cells is not a novel finding, nor is the role of IL10 in T cell dysfunction(2,3).  
+
+<|ref|>text<|/ref|><|det|>[[118, 471, 881, 551]]<|/det|>
+Indeed, this mechanism is reported in other cancer types but not for brain malignancy so far. Other cancer types can also be treated with checkpoint inhibitors, which is not possible for GBM. We think that investigating this special environment expands our comprehension. The fact that we found similar mechanism that can be also observed in other cancer types is not unexpected.  
+
+<|ref|>text<|/ref|><|det|>[[118, 574, 880, 612]]<|/det|>
+b. Using in vitro stimulated T cells to compare with in vivo T cells coming from a tumor ignores much of the complexity in signals that is occurring intratumorally.  
+
+<|ref|>text<|/ref|><|det|>[[118, 635, 880, 673]]<|/det|>
+The stimulation experiments are used to detect downstream pathway activation based on an isolated cytokine. We remove all other interpretations.  
+
+<|ref|>text<|/ref|><|det|>[[115, 696, 877, 714]]<|/det|>
+c. Using a second cell line in the in vitro neocortical studies would increase the impact of related findings.  
+
+<|ref|>text<|/ref|><|det|>[[118, 737, 880, 775]]<|/det|>
+Since this work does not focus on the tumor directly, using multiple donors to investigate the variance across patients was our main focus.  
+
+<|ref|>text<|/ref|><|det|>[[118, 799, 881, 899]]<|/det|>
+In general, authors show novel and interesting computational analyses from cutting edge techniques, however they lack substance in their follow up mechanistic studies. The use of pseudotime and RNA velocity to interrogate T cell activation states and pair them with GBM subtypes and spatial information are intriguing. Additionally, the nearest functionally connected neighbor algorithm will add to the expanding pool of resources for inferring intercellular communication from scRNAseq data. Follow- up
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[117, 83, 880, 140]]<|/det|>
+studies are limited in their support of computational findings (IL- 10 mediated T cell exhaustion); therefore, additional validation studies are required, or the focus of the story should shift to highlight the novel bioinformatic analyses.  
+
+<|ref|>text<|/ref|><|det|>[[117, 144, 880, 182]]<|/det|>
+Thank you for the appreciation of the computational results and tools. However, we think that biological validation, even if limited, supports the computational analysis.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[119, 83, 404, 99]]<|/det|>
+Reviewer #3 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 103, 882, 530]]<|/det|>
+Reviewer #3 (Remarks to the Author):Ravi, Neidert, Will et al present a single- cell RNA- sequencing study of tumor- infiltrating lymphocytes of patients with GBM. The profile 8 patients and additionally show data for 3 additional patients using spatial profiling, all using 10x platforms. Using established and novel analytical tools to infer trajectories of cell differentiation, they identify variability among T cells; among other, they find sub- clusters of CD8+ T cells with high expression of dysfunction marker TIM3 (HAVCR2) and cells with a hypoxia signature with distinct trajectories. They correlate these signatures with signatures of T cells collected following in vitro stimulation with different cytokines, including IL2, IFNG and IL10, arguing that IL10 stimulated cells have a lower activation sore compared to other cells identified as effector cells. Using spatial RNA- seq of three tumors, they find an association of tumor- mesenchymal niches and infiltration of exhausted T cells (TIM3/LAG3), suggesting that niches of dysfunctional T cells may be in part explained by cancer cell intrinsic features. In order to understand the origin of IL10 and to solidify the role of T cells as IL10 recipients, they present a analytical framework that infers ligand- receptor interactions using several constraints, and demonstrate that myeloid cells (CD163+, HMOX1+) are a main source of IL10. To begin validating this finding, they use slice cultures that they deplete of myeloid cells and show that depletion of myeloid cells results in reduction of IL10 (in the presence of tumor cells); in these models, they then co- culture autologous T cells and show that depletion of myeloid cells in slice cultures results in increased IL2, but not IFNG protein abundance. Incubation of T cells with an IL10R- inhibitor prior to co- culture with tissues results in increased IL2 production in T cells. Because the JAK/STAT pathway is downstream of IL10 signaling, they use ruxolitinib (a selective JAK1/2 inhibitor) first slice model showing increased IL2, and then use this drug in the neo- adjuvant therapy of a patient with GBM, followed by analysis of the surgical specimen, which shows activation of T cells.  
+
+<|ref|>text<|/ref|><|det|>[[118, 553, 881, 631]]<|/det|>
+GBM is a disease with extremely poor prognosis, and therapeutic development has in part been hampered by limited understanding of the tumor microenvironment; as such, the study of potentially high importance. However, several aspects raised significant concerns and reduced enthusiasm for this study, and need to be addressed,  
+
+<|ref|>text<|/ref|><|det|>[[118, 656, 880, 692]]<|/det|>
+We would like to thank the reviewer for his time and comments leading to an improvement of the manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[118, 718, 216, 733]]<|/det|>
+Major points:  
+
+<|ref|>text<|/ref|><|det|>[[117, 758, 881, 855]]<|/det|>
+1. Nowhere in the manuscript do the authors describe the characteristics of the patient tumors used for either single-cell sequencing of spatial sequencing. Are these all treatment naive tumors? where they exposed to different therapies (radiation, chemotherapy, immunotherapy, investigational drugs) - this will have a dramatic impact on the measured T cell phenotypes and in and of itself could describe variability seen in the data set.  
+
+<|ref|>text<|/ref|><|det|>[[118, 860, 880, 897]]<|/det|>
+All samples used for the dataset are naive non-treated primary GBM samples except the JAK-inhibitor treated samples as described in the last section on the manuscript.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[118, 82, 881, 161]]<|/det|>
+The authors should describe basic demographics and treatment history; it is not reasonable to request from the authors to attempt to account for variability based on basic demographics (this, and virtually any single- cell study would be underpowered), but they should show major analyses/findings in the context of different therapies received to exclude the possibility noted above.  
+
+<|ref|>text<|/ref|><|det|>[[118, 185, 880, 223]]<|/det|>
+Indeed, prior treatment can strongly affect the immune compartment. Here, only primary non- treated samples are included. We have added a supplementary table for demographic details.  
+
+<|ref|>text<|/ref|><|det|>[[117, 247, 882, 427]]<|/det|>
+2. Technical quality: it is somewhat surprising that the authors only recover \(\sim 1000\) unique genes per cell with only \(\sim 2300\) unique molecular identifiers - this is not on par with the quality described and raises concerns regarding data quality; this is particularly surprising as they used the 3.1 chemistry which performs better than prior chemistries; in fact recent studies performing profiling from frozen tissues even achieved similar or better quality compared to this present study (Slyper, Nature Medicine, 2020). Indeed, the sequencing depth of the samples was only moderate (Sequencing Saturation \(\sim 10\% - 15\%\) ) and we decided to re-sequence all our libraries. We gained an improvement in quality to approximately \(\sim 2400\) genes per cell and \(\sim 10k\) unique molecular identifiers. To provide an overview of the dataset quality, we have added some comparisons to recent published datasets, Supplementary Figure 2.  
+
+<|ref|>text<|/ref|><|det|>[[117, 451, 881, 550]]<|/det|>
+Furthermore, some of the clusters described might be artifactual due to tissue processing (e.g. "hypoxia cluster")- this possibility should be addressed using available data sets systematically investigating such artifacts (e.g. Ido Amit laboratory). Furthermore, the authors should comment on the technical quality. An additional embedding showing the UMI count for major analyses should be shown in the supplement to exclude the possibility of technical artifacts as drivers of clustering.  
+
+<|ref|>text<|/ref|><|det|>[[117, 574, 881, 652]]<|/det|>
+We added a scRNA- seq quality check in the supplementary results. Using a recently reported cell- type alignment algorithm (WNN<sup>3</sup>), we redesigned the first part and specifically investigated CD8 and CD4 T cells separately. As recommended, we have opted for alignment to reference datasets. The stress cluster within the T cell population has already been confirmed in the meantime in another cohort<sup>4</sup>.  
+
+<|ref|>text<|/ref|><|det|>[[117, 676, 882, 816]]<|/det|>
+3. There is no statistical evaluation of the inference made in Figure 3f - the authors should provide this; in this same figure, they also show that CCL2 myeloid cells are scoring highly, which is a gene considered to be an immunostimulatory gene/protein - how do they reconcile this? This brings up the question about a more nuanced annotation of the myeloid cells beyond monocytes, macrophages and microglia. The effect size of the gene set enrichment analysis in 3h is very underwhelming. In fact, throughout this section and the studies shown in Figure 4, the effect sizes are very small with in part borderline significance, raising the question of biological significance of these findings.  
+
+<|ref|>text<|/ref|><|det|>[[118, 840, 881, 899]]<|/det|>
+We have revised large parts of the results presented formerly in Figure 4. This was done by resequencing and improving the vertical integration, have eliminated the previously seen myeloid cell populations. As already mentioned above, the myeloid populations seem to have been an artefact of
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[118, 82, 880, 140]]<|/det|>
+insufficient vertical integration. Retrospectively, the integration used at that time by means of the cell ranger pipeline was not a sufficient approach. The MNN integration used now offers a much better approach to data integration.  
+
+<|ref|>text<|/ref|><|det|>[[118, 164, 881, 222]]<|/det|>
+4. The experiments in slice cultures should be described in more detail in the main text. Here, they state that they performed "myeloid depletion" when in fact they performed microglia depletion (as stated as header in the methods section).  
+
+<|ref|>text<|/ref|><|det|>[[118, 226, 881, 285]]<|/det|>
+We have tried to discuss the experimental part, especially the limitations, in detail. Our new data and analyses better support the reported results. The chemical depletion of microglia used here, the myeloid population in the human brain slice, has already been described in detail our previous work<sup>1,2</sup>.  
+
+<|ref|>text<|/ref|><|det|>[[118, 308, 881, 366]]<|/det|>
+The effect size of myeloid/microglia depletion on IL10 production is rather modest. One missing control is depletion of other cell types within the slice culture and measurement of the effect on IL10 to exclude the possibility that there are other major sources of IL10 production (which is likely).  
+
+<|ref|>text<|/ref|><|det|>[[117, 370, 882, 510]]<|/det|>
+Indeed, we agree that there are other sources of IL10 release in the tumor microenvironment. Neurons, astrocytes, and oligodendrocytes are potential candidates for IL10 release, but protocols for depletion of these cell types have not yet been established. To address this question, we attempted to establish cell- specific depletion for astrocytes. Unfortunately, we failed with this because the toxicity of astrocyte depletion is too high. Our results with IL10 inhibition confirm the downstream mechanism but cannot conclusively resolve the question of which cell type should also be considered as an IL10 source. At least our data suggest that a large fraction of IL10 is derived from HMOX1- positive myeloid cells.  
+
+<|ref|>text<|/ref|><|det|>[[118, 533, 881, 591]]<|/det|>
+Furthermore, it is surprising that IL2 production, but not IFNG production increases during T cell depletion - the authors should offer potential explanations as this argues against reinvigoration of T cell poly- functionality.  
+
+<|ref|>text<|/ref|><|det|>[[117, 595, 882, 734]]<|/det|>
+We agree with the reviewer that the IFNG signal is difficult to explain. When we examined the raw signal, we found generally high levels (including the negative control) of IFNG and other cytokines in the ELISA, suggesting a potential technical problem. We fully reviewed the ELISA data and re- ran the assay. The new data provided a clearer picture. Also in the new data, we see only a small baseline effect on IL10 after depletion of microglia. We assume that within our slice model the reactive transformation of microglia is necessary to upregulate IL10 release. Therefore, differences in the "no- tumor" slices were not observed.  
+
+<|ref|>text<|/ref|><|det|>[[118, 758, 881, 816]]<|/det|>
+The results shown in 4j confirm that IL10 is an immunosuppressive cytokine, but do not substantiate claims that this is mediated by myeloid cells. Again, perplexing that no change in interferon gamma is seen.  
+
+<|ref|>text<|/ref|><|det|>[[118, 820, 880, 858]]<|/det|>
+As mentioned above, we are limited within our model to elucidate all potential sources of IL10 release. Our data confirms that a significant part originates from myeloid cells.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[117, 83, 881, 140]]<|/det|>
+The single patient study is encouraging - was this pre-/post- comparison performed after single- agent therapy with ruxolitinib or was there a combination used? If the latter, it is possible that observed effects are due to other treatment constituents (see comment 1).  
+
+<|ref|>text<|/ref|><|det|>[[117, 144, 882, 264]]<|/det|>
+We used monotherapy in a neoadjuvant setting. However, the patient was pretreated with RT+TMZ +CCNU (CeTeG protocol) and received TTF (not in parallel to ruxolitinib). Resection of the tumor was subsequently performed after 6 weeks of ruxolitinib monotherapy. For analysis, we were able to perform staining for direct pre/post- treatment. Unfortunately, single cell sequencing could not be performed from the very small biopsy sample (before therapy), but only from the surgery sample (after treatment). We added a more detailed description and illustrations in the updated manuscript.
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[118, 84, 213, 99]]<|/det|>
+## Bibliography  
+
+<|ref|>text<|/ref|><|det|>[[117, 117, 868, 214]]<|/det|>
+1. Ravi, V. M. et al. Human organotypic brain slice culture: a novel framework for environmental research in neuro-oncology. Life Sci. Alliance 2, (2019).  
+2. Henrik Heiland, D. et al. Tumor-associated reactive astrocytes aid the evolution of immunosuppressive environment in glioblastoma. Nat. Commun. 10, 2541 (2019).  
+3. Hao, Y., Hao, S. & Andersen, E. Integrated analysis of multimodal single-cell data. Nissen  
+4. Mathewson, N. D. et al. Inhibitory CD161 receptor identified in glioma-infiltrating T cells by single-cell analysis. Cell 184, 1281–1298.e26 (2021).
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[119, 83, 312, 98]]<|/det|>
+## REVIEWER COMMENTS  
+
+<|ref|>text<|/ref|><|det|>[[120, 113, 402, 128]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[119, 142, 865, 187]]<|/det|>
+The authors have substantially improved the presentation of the hypotheses, analyses, and data in the revised manuscript, and have incorporated several new analyses that fill some of the gaps in the previous version. In my opinion, the revised manuscript is suited for publication.  
+
+<|ref|>text<|/ref|><|det|>[[117, 201, 832, 216]]<|/det|>
+I would only like to point some typos and small suggestions to improve the clarity of some parts:  
+
+<|ref|>text<|/ref|><|det|>[[117, 230, 580, 394]]<|/det|>
+- Line 123: "reference datasets" -> "reference dataset"  
+- Line 137: "Supplementary Figure 1b-c" -> "Figure 1b-c"  
+- Line 189: "Figure 2b" -> "Figure 2c"  
+- Line 223: "Figure 3d" -> "Figure 3f"  
+- Line 309: "Figure 5b" -> "Figure 5c"  
+- Line 389: "Benjamini-Hochberger" -> "Benjamini-Hochberg"  
+
+<|ref|>text<|/ref|><|det|>[[118, 407, 875, 437]]<|/det|>
+- Line 316: the notation used for the two clusters of myeloid cells (aM\Phi and bM\Phi) does not match with the notation used in Supplementary Fig. 5.  
+
+<|ref|>text<|/ref|><|det|>[[118, 451, 872, 511]]<|/det|>
+- In figure 1d,e it is unclear what the authors mean by "z-scored Gene Expression". z-scores are not bounded between 0 and 1, so I suspect the authors refer to something else or they have rescaled the z-scores in some way to lie between 0 and 1. I would suggest clarifying the normalization used in the figure legend and in the methods section.  
+
+<|ref|>text<|/ref|><|det|>[[118, 525, 866, 540]]<|/det|>
+- The specific gene sets from MSigDB v7 that were used should be specified in the methods section.  
+
+<|ref|>text<|/ref|><|det|>[[118, 555, 875, 599]]<|/det|>
+- I find the notation used in Fig. 5c and other figures (circle color + size) to be confusing. For example, in Fig. 5c the Mes-like correlation of #URK_S3 seems larger than the Mes-like correlation of #UKF_S2 based on the color of the circle, but smaller or equal based on the size of the circle.  
+
+<|ref|>text<|/ref|><|det|>[[118, 614, 866, 644]]<|/det|>
+- It would be helpful to add more details in the legend of Fig. 5b. For example, what does each of the colors in the figure denotes?  
+
+<|ref|>text<|/ref|><|det|>[[120, 673, 402, 687]]<|/det|>
+Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[118, 702, 868, 850]]<|/det|>
+In their revised manuscript titled "Crosstalk between lymphoid and myeloid cells orchestrates glioblastoma immunity through Interleukin 10 signaling", investigators coupled scRNAseq and stRNAseq to query the tumor microenvironment of 8 treatment naive glioblastoma patients. Ligand/receptor interactions were identified using a novel algorithm, nearest functionally connected neighbors (NFCN), ultimately identifying HMOX1+ myeloid cells as a major source of IL- 10. A T cell exhaustion phenotype was linked to the HMOX1+ myeloid cells and validated with both stRNAseq and an ex vivo neocortical system. In this revision, authors addressed concerns regarding the limitations of ex vivo experiments in validation of in silico findings. Additionally, authors increased the detail in many of the results sections to clarify relevant findings. These revisions greatly improve the quality and impact of the manuscript.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[118, 98, 402, 113]]<|/det|>
+Reviewer #3 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[118, 127, 855, 158]]<|/det|>
+The authors have done a very good job revising the manuscript and addressing my comments and suggestions.  
+
+<|ref|>text<|/ref|><|det|>[[118, 171, 867, 202]]<|/det|>
+I would strongly encourage them to highlight some of the technical and experimental challenges they had in their revision, as this might be of importance for future studies by their and other groups.  
+
+<|ref|>text<|/ref|><|det|>[[118, 216, 223, 231]]<|/det|>
+Benjamin Izar
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[119, 85, 580, 112]]<|/det|>
+## Point - by - Point Revision 2  
+
+<|ref|>text<|/ref|><|det|>[[119, 124, 222, 139]]<|/det|>
+D. H. Heiland  
+
+<|ref|>text<|/ref|><|det|>[[118, 165, 880, 202]]<|/det|>
+We thank all reviewers for their time and effort in evaluating our manuscript and appreciate the positive feedback on our project.  
+
+<|ref|>text<|/ref|><|det|>[[119, 226, 402, 242]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[118, 266, 881, 325]]<|/det|>
+The authors have substantially improved the presentation of the hypotheses, analyses, and data in the revised manuscript, and have incorporated several new analyses that fill some of the gaps in the previous version. In my opinion, the revised manuscript is suited for publication.  
+
+<|ref|>text<|/ref|><|det|>[[118, 348, 832, 366]]<|/det|>
+I would only like to point some typos and small suggestions to improve the clarity of some parts:  
+
+<|ref|>text<|/ref|><|det|>[[118, 388, 525, 405]]<|/det|>
+- Line 123: "reference datasets" -> "reference dataset"  
+
+<|ref|>text<|/ref|><|det|>[[118, 410, 191, 425]]<|/det|>
+Changed  
+
+<|ref|>text<|/ref|><|det|>[[118, 430, 540, 446]]<|/det|>
+- Line 137: "Supplementary Figure 1b-c" -> "Figure 1b-c"  
+
+<|ref|>text<|/ref|><|det|>[[118, 452, 191, 466]]<|/det|>
+Changed  
+
+<|ref|>text<|/ref|><|det|>[[118, 471, 392, 487]]<|/det|>
+- Line 189: "Figure 2b" -> "Figure 2c"  
+
+<|ref|>text<|/ref|><|det|>[[118, 492, 191, 507]]<|/det|>
+Changed  
+
+<|ref|>text<|/ref|><|det|>[[118, 512, 390, 528]]<|/det|>
+- Line 223: "Figure 3d" -> "Figure 3f"  
+
+<|ref|>text<|/ref|><|det|>[[118, 533, 191, 548]]<|/det|>
+Changed  
+
+<|ref|>text<|/ref|><|det|>[[118, 553, 392, 568]]<|/det|>
+- Line 309: "Figure 5b" -> "Figure 5c"  
+
+<|ref|>text<|/ref|><|det|>[[118, 574, 191, 589]]<|/det|>
+Changed  
+
+<|ref|>text<|/ref|><|det|>[[118, 593, 576, 610]]<|/det|>
+- Line 389: "Benjamini- Hochberger" -> "Benjamini-Hochberg"  
+
+<|ref|>text<|/ref|><|det|>[[118, 615, 191, 630]]<|/det|>
+Changed  
+
+<|ref|>text<|/ref|><|det|>[[118, 635, 880, 672]]<|/det|>
+- Line 316: the notation used for the two clusters of myeloid cells (aM\Phi and bM\Phi) does not match with the notation used in Supplementary Fig. 5.  
+
+<|ref|>text<|/ref|><|det|>[[118, 676, 412, 692]]<|/det|>
+Changed in the new supplementary file  
+
+<|ref|>text<|/ref|><|det|>[[117, 696, 881, 775]]<|/det|>
+- In figure 1d,e it is unclear what the authors mean by "z-scored Gene Expression". z-scores are not bounded between 0 and 1, so I suspect the authors refer to something else or they have rescaled the z-scores in some way to lie between 0 and 1. I would suggest clarifying the normalization used in the figure legend and in the methods section.  
+
+<|ref|>text<|/ref|><|det|>[[118, 779, 624, 796]]<|/det|>
+It is normalized gene expression, we changed the figure description  
+
+<|ref|>text<|/ref|><|det|>[[118, 799, 866, 816]]<|/det|>
+- The specific gene sets from MSigDB v7 that were used should be specified in the methods section.  
+
+<|ref|>text<|/ref|><|det|>[[118, 821, 500, 836]]<|/det|>
+The gene sets are implemented in the method part.  
+
+<|ref|>text<|/ref|><|det|>[[117, 840, 881, 898]]<|/det|>
+- I find the notation used in Fig. 5c and other figures (circle color + size) to be confusing. For example, in Fig. 5c the Mes-like correlation of #URK_S3 seems larger than the Mes-like correlation of #UKF_S2 based on the color of the circle, but smaller or equal based on the size of the circle.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[117, 83, 694, 98]]<|/det|>
+It was based on the fact that plots were created individually, we changed this.  
+
+<|ref|>text<|/ref|><|det|>[[117, 103, 880, 140]]<|/det|>
+- It would be helpful to add more details in the legend of Fig. 5b. For example, what does each of the colors in the figure denotes?  
+
+<|ref|>text<|/ref|><|det|>[[118, 145, 394, 160]]<|/det|>
+We added a description in the figure.  
+
+<|ref|>text<|/ref|><|det|>[[118, 186, 404, 201]]<|/det|>
+Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[116, 225, 882, 427]]<|/det|>
+In their revised manuscript titled "Crosstalk between lymphoid and myeloid cells orchestrates glioblastoma immunity through Interleukin 10 signaling", investigators coupled scRNAseq and stRNAseq to query the tumor microenvironment of 8 treatment naive glioblastoma patients. Ligand/receptor interactions were identified using a novel algorithm, nearest functionally connected neighbors (NFCN), ultimately identifying HMOX1+ myeloid cells as a major source of IL- 10. A T cell exhaustion phenotype was linked to the HMOX1+ myeloid cells and validated with both stRNAseq and an ex vivo neocortical system. In this revision, authors addressed concerns regarding the limitations of ex vivo experiments in validation of in silico findings. Additionally, authors increased the detail in many of the results sections to clarify relevant findings. These revisions greatly improve the quality and impact of the manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[118, 472, 404, 488]]<|/det|>
+Reviewer #3 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[117, 512, 881, 591]]<|/det|>
+The authors have done a very good job revising the manuscript and addressing my comments and suggestions. I would strongly encourage them to highlight some of the technical and experimental challenges they had in their revision, as this might be of importance for future studies by their and other groups.  
+
+<|ref|>text<|/ref|><|det|>[[117, 615, 881, 714]]<|/det|>
+R2 & R3: Thank you for the positive response. We are pleased that the reviewers appreciated the considerable effort and the change during the revision process. In our opinion, a description of these changes in the discussion clearly hampers the red- thread and takes the focus away from the scientific results. In our opinion, the detailed process can be reconstructed very well from published reviewers' comments, so we have refrained from making a statement in the discussion part.
+
+<--- Page Split --->

peer_reviews/126ed42b3fd6d9a339d2cd64a17c61348e2a1fa4df128380a49ace31df8377ca/supplementary_0_Transparent Peer Review file/images_list.json

@@ -0,0 +1 @@
+[]

peer_reviews/126ed42b3fd6d9a339d2cd64a17c61348e2a1fa4df128380a49ace31df8377ca/supplementary_0_Transparent Peer Review file/supplementary_0_Transparent Peer Review file.mmd

@@ -0,0 +1,56 @@
+
+# nature portfolio  
+
+Peer Review File  
+
+# Mechanistic evaluation of enhanced graphene toxicity to Bacillus induced by humic acid adsorption  
+
+Corresponding Author: Professor Qing Zhao  
+
+This manuscript has been previously reviewed at another journal. This document only contains information relating to versions considered at Nature Communications.  
+
+This file contains all reviewer reports in order by version, followed by all author rebuttals in order by version.  
+
+Version 0:  
+
+Reviewer comments:  
+
+Reviewer #1  
+
+(Remarks to the Author) The authors have made proper revisions.  
+
+Reviewer #2  
+
+(Remarks to the Author)  
+
+The manuscript by Zhao et al. describes a comprehensive investigation of the effect of the humic acid corona on the toxicity of graphene nanosheets (GNSs) to Bacillus tropicus. I previously reviewed this manuscript and my initial impressions were that the study was thorough and the results were interesting and had potential for high impact. Following several rounds of revision, I believe the authors have addressed the major and minor concerns raised by the reviewers and significantly improved the clarity of the manuscript. I have no further suggestions for revision and believe the manuscript is now publishable in its present form.
+
+<--- Page Split --->
+
+Open Access This Peer Review File is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.  
+
+In cases where reviewers are anonymous, credit should be given to 'Anonymous Referee' and the source.  
+
+The images or other third party material in this Peer Review File are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.  
+
+To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/
+
+<--- Page Split --->
+
+We are delighted with the generally positive feedback regarding the manuscript. Please find below a list of our detailed responses to all comments.  
+
+## REVIEWER COMMENTS  
+
+Reviewer #1 (Remarks to the Author):  
+
+The authors have made proper revisions.  
+
+Thank you for providing us so many insightful comments and suggestions, which have significantly enhanced the quality of our manuscript.  
+
+Reviewer #2 (Remarks to the Author):  
+
+The manuscript by Zhao et al. describes a comprehensive investigation of the effect of the humic acid corona on the toxicity of graphene nanosheets (GNSs) to Bacillus tropicus. I previously reviewed this manuscript and my initial impressions were that the study was thorough and the results were interesting and had potential for high impact. Following several rounds of revision, I believe the authors have addressed the major and minor concerns raised by the reviewers and significantly improved the clarity of the manuscript. I have no further suggestions for revision and believe the manuscript is now publishable in its present form.  
+
+Thank you for the positive feedback. Your expert insights and suggestions are immensely valuable in enhancing the quality of the manuscript.
+
+<--- Page Split --->

peer_reviews/126ed42b3fd6d9a339d2cd64a17c61348e2a1fa4df128380a49ace31df8377ca/supplementary_0_Transparent Peer Review file/supplementary_0_Transparent Peer Review file_det.mmd

@@ -0,0 +1,78 @@
+<|ref|>title<|/ref|><|det|>[[73, 50, 295, 79]]<|/det|>
+# nature portfolio  
+
+<|ref|>text<|/ref|><|det|>[[75, 96, 296, 118]]<|/det|>
+Peer Review File  
+
+<|ref|>title<|/ref|><|det|>[[74, 160, 855, 211]]<|/det|>
+# Mechanistic evaluation of enhanced graphene toxicity to Bacillus induced by humic acid adsorption  
+
+<|ref|>text<|/ref|><|det|>[[74, 224, 455, 241]]<|/det|>
+Corresponding Author: Professor Qing Zhao  
+
+<|ref|>text<|/ref|><|det|>[[73, 275, 874, 302]]<|/det|>
+This manuscript has been previously reviewed at another journal. This document only contains information relating to versions considered at Nature Communications.  
+
+<|ref|>text<|/ref|><|det|>[[70, 313, 866, 329]]<|/det|>
+This file contains all reviewer reports in order by version, followed by all author rebuttals in order by version.  
+
+<|ref|>text<|/ref|><|det|>[[73, 366, 145, 379]]<|/det|>
+Version 0:  
+
+<|ref|>text<|/ref|><|det|>[[73, 392, 219, 405]]<|/det|>
+Reviewer comments:  
+
+<|ref|>text<|/ref|><|det|>[[73, 418, 160, 431]]<|/det|>
+Reviewer #1  
+
+<|ref|>text<|/ref|><|det|>[[73, 444, 355, 471]]<|/det|>
+(Remarks to the Author) The authors have made proper revisions.  
+
+<|ref|>text<|/ref|><|det|>[[73, 483, 162, 496]]<|/det|>
+Reviewer #2  
+
+<|ref|>text<|/ref|><|det|>[[73, 509, 238, 522]]<|/det|>
+(Remarks to the Author)  
+
+<|ref|>text<|/ref|><|det|>[[72, 521, 914, 602]]<|/det|>
+The manuscript by Zhao et al. describes a comprehensive investigation of the effect of the humic acid corona on the toxicity of graphene nanosheets (GNSs) to Bacillus tropicus. I previously reviewed this manuscript and my initial impressions were that the study was thorough and the results were interesting and had potential for high impact. Following several rounds of revision, I believe the authors have addressed the major and minor concerns raised by the reviewers and significantly improved the clarity of the manuscript. I have no further suggestions for revision and believe the manuscript is now publishable in its present form.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[72, 45, 916, 99]]<|/det|>
+Open Access This Peer Review File is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.  
+
+<|ref|>text<|/ref|><|det|>[[72, 99, 796, 113]]<|/det|>
+In cases where reviewers are anonymous, credit should be given to 'Anonymous Referee' and the source.  
+
+<|ref|>text<|/ref|><|det|>[[72, 112, 910, 165]]<|/det|>
+The images or other third party material in this Peer Review File are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.  
+
+<|ref|>text<|/ref|><|det|>[[72, 165, 618, 179]]<|/det|>
+To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[148, 89, 848, 135]]<|/det|>
+We are delighted with the generally positive feedback regarding the manuscript. Please find below a list of our detailed responses to all comments.  
+
+<|ref|>sub_title<|/ref|><|det|>[[149, 247, 390, 265]]<|/det|>
+## REVIEWER COMMENTS  
+
+<|ref|>text<|/ref|><|det|>[[149, 303, 460, 320]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[149, 358, 479, 375]]<|/det|>
+The authors have made proper revisions.  
+
+<|ref|>text<|/ref|><|det|>[[148, 386, 850, 432]]<|/det|>
+Thank you for providing us so many insightful comments and suggestions, which have significantly enhanced the quality of our manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[148, 469, 460, 486]]<|/det|>
+Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[147, 523, 852, 738]]<|/det|>
+The manuscript by Zhao et al. describes a comprehensive investigation of the effect of the humic acid corona on the toxicity of graphene nanosheets (GNSs) to Bacillus tropicus. I previously reviewed this manuscript and my initial impressions were that the study was thorough and the results were interesting and had potential for high impact. Following several rounds of revision, I believe the authors have addressed the major and minor concerns raised by the reviewers and significantly improved the clarity of the manuscript. I have no further suggestions for revision and believe the manuscript is now publishable in its present form.  
+
+<|ref|>text<|/ref|><|det|>[[148, 756, 850, 802]]<|/det|>
+Thank you for the positive feedback. Your expert insights and suggestions are immensely valuable in enhancing the quality of the manuscript.
+
+<--- Page Split --->

peer_reviews/127b4a1df348ee954b5fefb88eeed0676c975caa5b61439cfcd7368eac65b846/supplementary_0_Peer Review File/images_list.json

@@ -0,0 +1,25 @@
+[
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_0.jpg",
+    "caption": "Dynamic compression of liquid water at a compression rate (10 GPa/ms). The sample pressure over time is measured using simultaneously the \\(S\\vert \\mathrm{text}(SrB)_2[4]\\vert \\mathrm{text}(O)_2[7]\\vert \\mathrm{text}(Sm)^\\wedge (2 + )S\\) and the ruby fluorescence",
+    "footnote": [],
+    "bbox": [],
+    "page_idx": 0
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_1.jpg",
+    "caption": "Dynamic compression of liquid water at a compression rate (10 GPa/ms). The sample pressure over time is measured using simultaneously the S(Text(Si6)_[4]/text(O)_[7]/text(Sm)^(2+)/S and the ruby fluorescence",
+    "footnote": [],
+    "bbox": [
+      [
+        100,
+        90,
+        410,
+        280
+      ]
+    ],
+    "page_idx": 14
+  }
+]

peer_reviews/127b4a1df348ee954b5fefb88eeed0676c975caa5b61439cfcd7368eac65b846/supplementary_0_Peer Review File/supplementary_0_Peer Review File.mmd

@@ -0,0 +1,459 @@
+
+# nature portfolio  
+
+Peer Review File  
+
+Metastable water at several compression rates and its freezing kinetics into ice VII  
+
+![](images/Figure_unknown_0.jpg)
+  
+
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+
+Editorial note: Parts of this Peer Review File have been redacted as indicated to remove third- party material where no permission to publish could be obtained.  
+
+## REVIEWER COMMENTS  
+
+## Reviewer #1 (Remarks to the Author):  
+
+This work shows in situ, time resolved x- ray diffraction data on the H2O system subjected to a wide range of compression rates. The diffraction data show real- time structural response of the system and allow unambiguous identification of solid phases of ice. I think the presented data provide interesting and important insight into the freezing dynamics and help to resolve discrepancies from other measurements.  
+
+I had several thoughts on how the manuscript could be made clearer but, beyond these, would like to recommend publication.  
+
+1. Page 2 introduction: "having the simplest ice structure with..." I think the work "simplest" is subjective. Since the structure of ice VII is highly disordered, one could say an ordered phase, such as ice VIII is structurally simpler. More appropriate would be a specific statement like "simplest crystallographic unit cell".  
+
+2. Page 2 introduction: "is a topic of current focus" when stating this, I think citations should be provided.  
+
+3. Page 5 instrumentation: It is mentioned that the pressure is measured via multiple approaches. Hence, how does the reader know which is refered to when pressures are quoted throughout the manuscript. E.g. what is shown in Figure 1?  
+
+4. Page 5 pressure response of...compression ramps": "the convolution with" I'm not sure this effect is a convolution in the mathematical sense.  
+
+5. Page 5 pressure response of...compression ramps": "1.56 GPa" No quantified uncertainty is given for this number and, on the basis of Figure 2 where error bar are shown to be ±0.05 this number is stated too precisely. Assuming the error bars indicate the magnitude of a standard deviation, quoting the pressure to one decimal place is sufficient.  
+
+6. Page 8 Freezing pressure of...compression rate: "0.056, which matches...0.069" Use of the
+
+<--- Page Split --->
+
+word 'matches' requires some information on experimental uncertainty, which is not provided.  
+
+7. Page 8 Freezing pressure of..compression rate: "When metastable water freeze under a compression rate of 110 GPa/ms, it corresponds to a cooling rate of \(10^{\wedge}7 \text{K / s}\) ". I have no idea how this equivalence was derived, can the authors explain their argument?  
+
+8. Figure 2: I was confused by the meaning of the coloured horizontal bands indicating phase regions. Are these taken from other work? Or if they are they guides to the eye to indicate behaviour observed in this study, I found it confusing that they are horizontal and not vertical as each phase is only present for bounded periods of time.  
+
+9. Figure 4: How can pressure be inferred from this plot? Is it possible to show in some way? 10. Figure 4: Out of curiosity, is some interpolation or smoothing applied to the 2d 2theta vs time data?  
+
+11. Figure 4: "A mixture of ice VI and ice VII is still observed" I wondered if it is possible by Rietveld analysis (likely with a highly constrained model) to determine the phase fraction of VI and VII? From the 2d image, it indeed looks like the intensity of the ice VII 110 decreases as the ice VI forms. If it could be quantitatively shown that this change is due to direct transformation of the volume of ice VII in the x-ray beam, that would strengthen the interpretation given.  
+
+## Reviewer #2 (Remarks to the Author):  
+
+This manuscript describes the effects of rapid compression of liquid water to its crystalline phases using dynamic- piezo- Diamond- Anvil- Cell. The main goal of this study was to provide experimental details and perhaps a physical understanding the structural changes of water undergoing rapid temperature and/or pressure variation.  
+
+Although there have been many studies of phase transitions in both water and ice, this study focused on one particular phase transition where there have been different previously reported results and interpretations. This is a transition between metastable liquid water and particular dense crystalline phases. The authors of this study employed additional an experimental technique of time- resolved X- ray diffraction with a dynamic- piezo- Diamond- Anvil- Cell. The employment of this technique provided sufficient new data to resolve the
+
+<--- Page Split --->
+
+contradictions that are in the literature.  
+
+This is a very well written manuscript and I therefore have only a small list of comments that should be addressed in more detail.1) On page 3 the authors indicate that they are investigating whether ice VII is still the freezing state of metastable water?? Under what, P conditions are they referring to?2) On page 5 the authors give a value of 1.56 GPa? How accurate is this number? T, P effects?3) On page 6: It was stated and suggested that crystallization occurs into ice VII, although without time-resolved XRD measurements this proposition remains to be proven. Ref?4) Pg. 6: excellent agreement between the different pressure determinations. Can the authors be clearer on this point? Perhaps by simply referring to the figures.5) On Pg. 7, it is stated that there was a e previous domain of investigation in d- DAC experiments. A reference. is needed here.6) without time-resolved XRD measurements this proposition remains to be proven. A reference should be made here.7) On page. 9 the authors state "Since water and ice are quite incompressible" Here a comparison with other triatomic molecules would strengthen this statement.8) It is important to emphasize that the model employed to fit the data is a classical nucleation theory description of the experimental results. It fits the data very well and provides a very good physical picture and interpretation of the experimental results but this depends on parameters used for fitting the data so small changes in the parameters employed could have a significant effect on the model used. This could be mentioned by the authors.  
+
+This report follows up on the previous domain of investigation in d- DAC experiments focusing on the solidification of metastable water. This is the key goal reached in this study. The authors clearly state that support the hypothesis that metastable water, within the pressure range of 1.5 GPa to 2.1 GPa, nucleates into ice VII first, despite this range being the stability field of ice VI. results offer a clarification to the apparent contradiction of the
+
+<--- Page Split --->
+
+previous findings.  
+
+previous findings.In summary, this manuscript clearly addresses and presents a solution to the phase transition discrepancies in the literature. A key improvement in the measurement technique was time- resolved X- ray diffraction. This technique is very well described in the manuscript All figures were clearly presented and definitely provided added clarity to the report. Acceptance is therefore suggested after the minor comments listed above are addressed.  
+
+## Reviewer #3 (Remarks to the Author):  
+
+The authors investigated the interesting behaviour of water ice under dynamic compression with various compression rates using d- DAC by time- resolved x- ray diffraction method and found that ice VII crystallise under the conditions where ice VI should be stable. They also interpret the excess pressure and growth time during the crystallization of ice VII using a phenomenological model based on the classical nucleation theory, succeeding in a comprehensive explanation for the growth rate of ice VII grown under various compression rates. The investigated technique, x- ray diffraction with ruby fluorescence observations at micro- to milli- second order, is somehow remarkable, and the overall achievements of this study would be worth recognising for the ice community. Thus, I do not doubt that this manuscript should be eventually published in at least some specific journal, but I am not sure whether this manuscript will be published by Nature Communication.  
+
+Two key points in assessing the value of this study would be 1) how new the observed results are and 2) how much their interpretation will influence other related studies. First, concerning the observed results, it should be stated that no truly new phenomena have been reported in this study. It has long been known among ice researchers that ice VII nucleates in the stable region of ice VI even under static compression (e.g. K. Yamamoto, Jpn J Appl Phys, 19, 1841, 1980, doi: 10.1143/JJAP.19.1841). A recent study using d- DAC has also reported ice VII nucleation in the stability region of ice VI [18]. Therefore, what is new in this manuscript are the results of quantitative experiments on how the growth rate of ice VII changes when the compression rate is varied. The experimental results in this study are novel in the sense that they are quantitative, but the observed phenomena themselves are not particularly novel. The manuscript also does not elaborate on how the quantitative growth rates obtained might affect other related studies. It is therefore difficult to assess, at
+
+<--- Page Split --->
+
+least at the moment, how important these results are.  
+
+Even if the importance of this manuscript is recognised by other reviewers and considered for publication in Nat. Commun., the following concerns remain and should be fully considered before potential publication.  
+
+- The authors assume that the sample temperature should not be increased by the rapid compression, but it should be more carefully evaluated and explained in the manuscript. I also believe that temperature might be quickly stable at room temperature owing to the high thermal conductivity of the diamond. However, if the sample size is enough large, this might not be the case since the thermal conductivity of ice VII would not be such high, so the temperature could temporally increase in a very short time. This issue should be quantitatively assessed, for example, by FEM simulation.  
+
+- In the adopted phenomenological model for the nucleation rate, authors ignore the effect of crystal growth, but only take the nucleation rate into account. However, I recognized that the grown ice VII may be somehow coarse crystalline size because the observed diffraction patterns are not like smooth lines but spotty. In particular, I found 111 spots of ice VII, which should be very very weak (generally invisible) if the specimen is fine powder, showing the specimen should be somehow coarse aggregates of many single crystals. The actual rate of ice VII growth may be expressed as the combination of nucleation and growth rates. In fact, the results obtained in this study could be interpreted by different phenomenological models, such as the simple JMAK model. The point would be 'which model is the best' to describe the observed phenomena. Even if one model fits well, this does not necessarily mean that this is the only correct model.  
+
+Plus, I am also concerned about the lack of temperature terms in the adopted model. I suppose the temperature term should appear since the classical nucleation theory generally has it. Authors should discuss this point (why temperature term could be ignored), if they stick to this model only.  
+
+- Previous studies using d-DAC reported that ice VI and HDA coexist [19,20] as mentioned in the manuscript, but HDA was not observed in this study. The authors should describe the difference from the previous study.
+
+<--- Page Split --->
+
+## Minor points  
+
+1. The precise wavelength should be described in Method section. I found the x-ray energy is 19 keV, but the number of significant figures is only two in this case.  
+2. The reflection indices should not be in parenthesis. Index in parenthesis, (hkl) means Miller index that is used for corresponding Miller plane in real space. But reflection index, hkl, in reciprocal space should not have any parenthesis (see International Tables for crystallography, for this criterion).  
+3. Effective digit and/or the error should be considered throughout the whole manuscript, in particular, in Table I
+
+<--- Page Split --->
+
+## Detailed answer to reviewers' comments.  
+
+## Reviewer #1  
+
+This work shows in situ, time resolved \(x\) - ray diffraction data on the H2O system subjected to a wide range of compression rates. The diffraction data show real- time structural response of the system and allow unambiguous identification of solid phases of ice. I think the presented data provide interesting and important insight into the freezing dynamics and help to resolve discrepancies from other measurements.  
+
+I had several thoughts on how the manuscript could be made clearer but, beyond these, would like to recommend publication.  
+
+1. Page 2 introduction: "having the simplest ice structure with..." I think the work "simplest" is subjective. Since the structure of ice VII is highly disordered, one could say an ordered phase, such as ice VIII is structurally simpler. More appropriate would be a specific statement like "simplest crystallographic unit cell".  
+
+The sentence was modified accordingly.  
+
+2. Page 2 introduction: "is a topic of current focus" when stating this, I think citations should be provided.  
+
+We have added citations.  
+
+3. Page 5 instrumentation: It is mentioned that the pressure is measured via multiple approaches. Hence, how does the reader know which is referred to when pressures are quoted throughout the manuscript. E.g. what is shown in Figure 1?  
+
+As mentioned in the manuscript, the imaging and X- ray diffraction (XRD) processes could not be conducted simultaneously. The pressure was measured using luminescence during imaging and by determining the volume of copper during XRD.  
+
+An essential aspect to highlight, as illustrated in Figure 3, is the high degree of repeatability and reproducibility of the compression ramps using the different diagnostic techniques. To enhance clarity, we have included the following sentence:  
+
+"The pressure was determined using a luminescence gauge during imaging and by utilizing XRD equation- of- state (EOS) data during XRD. The results from the different pressure measurement methods were found to be highly consistent with each other."  
+
+4. Page 5 pressure response of...compression ramps": "the convolution with" I'm not sure this effect is a convolution in the mathematical sense.  
+
+The sentence now reads: "However, if a phase transition occurs, the pressure rise is disrupted by the pressure drop associated to the negative volume discontinuity at the transition, resulting in an apparent negative compressibility."
+
+<--- Page Split --->
+
+5. Page 5 pressure response of...compression ramps": "1.56 GPa" No quantified uncertainty is given for this number and, on the basis of Figure 2 where error bar are shown to be \(\pm 0.05\) this number is stated too precisely. Assuming the error bars indicate the magnitude of a standard deviation, quoting the pressure to one decimal place is sufficient.  
+
+The pressure was modified to 1.6 GPa to take into account the \(\pm 0.05\) GPa uncertainty.  
+
+6. Page 8 Freezing pressure of...compression rate: "0.056, which matches...0.069" Use of the word 'matches' requires some information on experimental uncertainty, which is not provided.  
+
+The sentence now reads: "Remarkably, the exponent 'c' obtained from the fit is 0.056, which closely aligns with the value of 0.069 predicted by Myint's scaling law."  
+
+7. Page 8 Freezing pressure of compression rate: "When metastable water freeze under a compression rate of 110 GPa/ms, it corresponds to a cooling rate of \(10^{7} \mathrm{~K} / \mathrm{s}\) ". I have no idea how this equivalence was derived, can the authors explain their argument.  
+
+At a compression rate of 110 GPa/ms at 300 K, the freezing process occurs at 2.9 GPa after an over-compression duration of 18 us in the metastable water state. To achieve the same state through a 18 us cooling at 2.9 GPa from the liquid state, considering the melting point of water at 2.9 GPa is 423 K (Datchi et al., PRB 61, 6535 (2000)), the required cooling rate would have to be \(7 \times 10^{6} \mathrm{~K} / \mathrm{s}\) , approximately \(10^{7} \mathrm{~K} / \mathrm{s}\) . This calculation has been included in the supplementary material.  
+
+8. Figure 2: I was confused by the meaning of the coloured horizontal bands indicating phase regions. Are these taken from other work? Or if they are they guides to the eye to indicate behaviour observed in this study, I found it confusing that they are horizontal and not vertical as each phase is only present for bounded periods of time.  
+
+The colored regions in the figure represent the (meta)stability pressure domains at 300 K for the various phases. These domains are inferred from the static phase diagram and the newly measured metastable melting of liquid- ice VII in this study. They are included as a visual aid to guide the interpretation of the data.  
+
+9. Figure 4: How can pressure be inferred from this plot? Is it possible to show in some way?  
+
+Pressure is inferred from the measured volumes of Sn, ice VII and ice VI. A reference to the Sn EoS has been added and a clarification has been added in the text which now reads: "Here, the ice sample remaining as a mixture of ice VI and ice VII is observed at 1.8 GPa, as inferred from the measured volumes of Sn, ice VII and VI".  
+
+10. Figure 4: Out of curiosity, is some interpolation or smoothing applied to the 2d 2theta vs time data?
+
+<--- Page Split --->
+
+A gourdash shading is indeed applied to the data.  
+
+11. Figure 4: "A mixture of ice VI and ice VII is still observed" I wondered if it is possible by Rietveld analysis (likely with a highly constrained model) to determine the phase fraction of VI and VII? From the 2d image, it indeed looks like the intensity of the ice VII 110 decreases as the ice VI forms. If it could be quantitatively shown that this change is due to direct transformation of the volume of ice VII in the x-ray beam, that would strengthen the interpretation given.  
+
+Although the presence of a mixture of ice VI and ice VII is undeniable, conducting a Rietveld analysis to determine the exact phase fractions of ice VI and ice VII appears impracticable here. This is mainly due to the fact that ice VI and ice VII exhibit different textures and the detector angle-coverage is only partial.  
+
+## Reviewer #2  
+
+This manuscript describes the effects of rapid compression of liquid water to its crystalline phases using dynamic- piezo- Diamond- Anvil- Cell. The main goal of this study was to provide experimental details and perhaps a physical understanding the structural changes of water undergoing rapid temperature and/or pressure variation.  
+
+Although there have been many studies of phase transitions in both water and ice, this study focused on one particular phase transition where there have been different previously reported results and interpretations. This is a transition between metastable liquid water and particular dense crystalline phases. The authors of this study employed additional an experimental technique of time- resolved X- ray diffraction with a dynamic- piezo- Diamond- Anvil- Cell. The employment of this technique provided sufficient new data to resolve the contradictions that are in the literature.  
+
+This is a very well written manuscript and I therefore have only a small list of comments that should be addressed in more detail.  
+
+1) On page 3 the authors indicate that they are investigating whether ice VII is still the freezing state of metastable water?? Under what, P conditions are they referring to?  
+
+The previous speculations about the transformation of water into ice VII during shock compression experiments have been uncertain due to the lack of X- ray diffraction diagnostic. In this study, we aim to clarify this phenomenon by investigating the crystallization of metastable water under high compression rates within the stability field of ice VII. Specifically, we seek to determine whether the freezing state of metastable water under these conditions is indeed ice VII or another metastable structure. The text has been slightly modified to make it clearer: 'In particular, we are investigating whether the freezing state of metastable water at high compression rates above 2 GPa is indeed ice VII or a metastable structure....'
+
+<--- Page Split --->
+
+2) On page 5 the authors give a value of1.56 GPa? How accurate is this number? T, P effects?  
+
+Referee #1 has also brought up a similar point. To address this, we have adjusted the pressure to 1.6 GPa, which accounts for the uncertainty of \(\pm 0.05\) GPa.  
+
+3) On page 6: It was stated and suggested that crystallization occurs into ice VII, although without time-resolved XRD measurements this proposition remains to be proven. Ref?  
+
+The relevant references (14- 17) have already been cited a few lines earlier. To avoid redundancy and maintain the flow of the text, we have chosen not to repeat them in this instance.  
+
+4) Pg. 6: excellent agreement between the different pressure determinations. Can the authors be clearer on this point? Perhaps by simply referring to the figures.  
+
+A reference to Figure 3 was added accordingly, showing that the pressure determination from the different gauges (Cu, Ice VII and SrB4O7 : Sm2+) are consistent with each other.  
+
+5) On Pg. 7, it is stated that there was a e previous domain of investigation in d-DAC experiments. A reference. is needed here.  
+
+The corresponding references (18- 20) have been added.  
+
+6) without time-resolved XRD measurements this proposition remains to be proven. A reference should be made here.  
+
+This point has been answered in point #3.  
+
+7) On page. 9 the authors state "Since water and ice are quite incompressible" Here a comparison with other triatomic molecules would strengthen this statement.  
+
+The bulk modulus of water ice is approximately 20 GPa, which is relatively high compared to other molecular systems, such as NH3 (4.2 GPa). However, the focus of this analysis is not on the incompressibility of the system, but rather on the assumption of constant compressibility. By using a first- order Birch equation of state (EoS), assuming constant compressibility, which is a reasonable approximation over the 10 GPa pressure range considered, we can derive a phenomenological model fit function with three free parameters (Equation 4). This function provides a good fit for the data obtained from diamond anvil cell (d- DAC), laser, and gas gun experiments, as shown in Figure 6.  
+
+We have removed the assertion: ' since water and ice are incompressible'.
+
+<--- Page Split --->
+
+8) It is important to emphasize that the model employed to fit the data is a classical nucleation theory description of the experimental results. It fits the data very well and provides a very good physical picture and interpretation of the experimental results but this depends on parameters used for fitting the data so small changes in the parameters employed could have a significant effect on the model used. This could be mentioned by the authors.  
+
+We have explicitly stated in the text that the phenomenological model used serves as a fitting function. To further clarify this, we have slightly modified the caption of Figure 6 to read: 'Fitted using the CNT- based phenomenological model given in Equation 4, with the three free parameters equal to ...  
+
+This report follows up on the previous domain of investigation in d- DAC experiments focusing on the solidification of metastable water. This is the key goal reached in this study. The authors clearly state that support the hypothesis that metastable water, within the pressure range of 1.5 GPa to 2.1 GPa, nucleates into ice VII first, despite this range being the stability field of ice VI. results offer a clarification to the apparent contradiction of the previous findings.  
+
+In summary, this manuscript clearly addresses and presents a solution to the phase transition discrepancies in the literature. A key improvement in the measurement technique was time- resolved X- ray diffraction. This technique is very well described in the manuscript All figures were clearly presented and definitely provided added clarity to the report. Acceptance is therefore suggested after the minor comments listed above are addressed.  
+
+## Reviewer #3  
+
+The authors investigated the interesting behaviour of water ice under dynamic compression with various compression rates using d- DAC by time- resolved x- ray diffraction method and found that ice VII crystallise under the conditions where ice VI should be stable. They also interpret the excess pressure and growth time during the crystallization of ice VII using a phenomenological model based on the classical nucleation theory, succeeding in a comprehensive explanation for the growth rate of ice VII grown under various compression rates. The investigated technique, x- ray diffraction with ruby fluorescence observations at micro- to milli- second order, is somehow remarkable, and the overall achievements of this study would be worth recognising for the ice community. Thus, I do not doubt that this manuscript should be eventually published in at least some specific journal, but I am not sure whether this manuscript will be published by Nature Communication. Two key points in assessing the value of this study would be 1) how new the observed results are and 2) how much their interpretation will influence other related studies. First, concerning the observed results, it should be stated that no truly new phenomena have been reported in this study. It has long been known among ice researchers that ice VII nucleates in the stable region of ice VI even under static compression (e.g. K. Yamamoto, Jpn J Appl Phys,
+
+<--- Page Split --->
+
+19, 1841, 1980, doi: 10.1143/JJAP.19.1841). A recent study using d- DAC has also reported ice VII nucleation in the stability region of ice VI [18]. Therefore, what is new in this manuscript are the results of quantitative experiments on how the growth rate of ice VII changes when the compression rate is varied. The experimental results in this study are novel in the sense that they are quantitative, but the observed phenomena themselves are not particularly novel. The manuscript also does not elaborate on how the quantitative growth rates obtained might affect other related studies. It is therefore difficult to assess, at least at the moment, how important these results are.  
+
+It is true that the nucleation of ice VII in the stability field of ice VI and the increase in the pressure of overcompressed metastable water with the compression rate were previously known phenomena. But being able to provide an explanation of phenomena through quantitative measurement is the ultimate goal of the physical science research.  
+
+As acknowledged by reviewer #3, the possibility to control the compression rate over 4 orders of magnitude and to perform time resolved XRD and imaging with the appropriate time- frame from microsecond to millisecond is a significant accomplishment. The quantitative measurements obtained through this approach have enabled us to clarify and unify all previous measurements of the freezing of overcompressed water in a coherent interpretation. Metastable water overcompression under dynamical compression has been the subject of many studies over the past 20 years, many published in high impact journals, as a textbook case of phase transition occurring under far- from- equilibrium conditions. In the present case of water solidification at extreme overcompression, we were able to test a universal scaling law, recently proposed. The Classical Nucleation Theory was shown to unify all previous measurements which lacked the microscopic XRD information.  
+
+Regarding the solidification of ice VII in the stability domain of ice VI, we thank reviewer#3 for providing this Jp. J. Appl. Phys. reference. However, in it, the reported observation of ice VII in the stability field of ice VI is not exactly what is reported in our study. The nucleation of ice VII was obtained at the metastable extension of the melting line of ice VII in the stability field of ice VI by heating the sample so that the overcompression of water that can be obtained under quasi- static compression could reach the metastable melting line of ice VII. Nucleation of ice VII is stated then to occasionally form and a facetted single crystal of ice VII can bestabilized in equilibrium with the liquid. However, ice VII is kept metastable only if facetted in equilibrium with the liquid. That is exactly what we observed and used in the experiment reported in our supplementary material to precisely extend the ice VII melting line in the stability domain of ice VI. We now have cited the reference in our supplementary information. However, out of this liquid- solid phase equilibrium, by slightly increasing pressure ice VII transforms into ice VI. We are reported a different story here. Under dynamical compression, metastable water always freeze into ice VII in the metastability domain of ice VII and then ice VII transforms to ice VI over a ms time scale and a mixture ice VI - ice VII then exists, which should eventually evolve to an amorphous - ice VI mixture. This resolves all previous contradictions of the literature between reports of amorphous HAD, HAD- ice VI mixture or ice VII.
+
+<--- Page Split --->
+
+Even if the importance of this manuscript is recognised by other reviewers and considered for publication in Nat. Commun., the following concerns remain and should be fully considered before potential publication.  
+
+- The authors assume that the sample temperature should not be increased by the rapid compression, but it should be more carefully evaluated and explained in the manuscript. I also believe that temperature might be quickly stable at room temperature owing to the high thermal conductivity of the diamond. However, if the sample size is enough large, this might not be the case since the thermal conductivity of ice VII would not be such high, so the temperature could temporally increase in a very short time. This issue should be quantitatively assessed, for example, by FEM simulation.  
+
+The temperature increase in the sample could be an issue. A recent d- DAC measurement leveraged this effect to measure the Gruneisen parameter but for doing that they had to put a layer of zirconia powder on the anvil tip to prevent heat loss from the diamond (ref. 25). We have taken steps to quantify the heating associated to rapid compression by using two pressure luminescence gauges, specifically ruby and SrB40.7:Sm2+, positioned inside the sample chamber, as detailed in ref. 29. The luminescence line of ruby depends on pressure and temperature, whereas that of Samarium depends only on pressure. This method has been previously used to accurately measure melting lines (Datchi et al, PRB 61, 6535). We thus could show that the heating of the sample is negligible. That is illustrated in the figure below showing that, no discernible heating was detected within the measurement's sensitivity, approximately \(\pm 3 \mathrm{~K}\) . Therefore, we can confidently assert that there is no heating occurring. We have now included this quantification of the absence of heating in the supplementary material.
+
+<--- Page Split --->
+![](images/Figure_unknown_1.jpg)
+
+<center>Dynamic compression of liquid water at a compression rate (10 GPa/ms). The sample pressure over time is measured using simultaneously the \(S\vert \mathrm{text}(SrB)_2[4]\vert \mathrm{text}(O)_2[7]\vert \mathrm{text}(Sm)^\wedge (2 + )S\) and the ruby fluorescence </center>  
+
+- In the adopted phenomenological model for the nucleation rate, authors ignore the effect of crystal growth, but only take the nucleation rate into account. However, I recognized that the grown ice VII may be somehow coarse crystalline size because the observed diffraction patterns are not like smooth lines but spotty. In particular, I found 111 spots of ice VII, which should be very very weak (generally invisible) if the specimen is fine powder, showing the specimen should be somehow coarse aggregates of many single crystals. The actual rate of ice VII growth may be expressed as the combination of nucleation and growth rates. In fact, the results obtained in this study could be interpreted by different phenomenological models, such as the simple JMAK model. The point would be 'which model is the best' to describe the observed phenomena. Even if one model fits well, this does not necessarily mean that this is the only correct model.  
+
+Plus, I am also concerned about the lack of temperature terms in the adopted model. I suppose the temperature term should appear since the classical nucleation theory generally has it. Authors should discuss this point (why temperature term could be ignored), if they stick to this model only.  
+
+First, it is important to note that there is no temperature term involved, as explained above. In the discussion, we begin with the JMAK model and make the assumption that the nucleation rate is predominant over the growth rate. This assumption has been previously utilized in the analysis of overcompression studies of water (ref. 15 and Ref. 16).  
+
+We have explicitly stated in the text that, for compression rates exceeding a few GPa/ms, this assumption appears reasonable. This conclusion is drawn from the imaging observation of instantaneous nucleation throughout the entire sample chamber, as well
+
+<--- Page Split --->
+
+as the absence of a plateau in the pressure - time evolution during compression. This pertains to the data points presented in figure 6. Conversely, for data points obtained below 1 GPa/ms, a plateau is observed and the growth of the crystallites is imaged, as shown in figure 2. So
+
+<--- Page Split --->
+
+, we believe we correctly explain the framework of our analysis.  
+
+- Previous studies using d-DAC reported that ice VI and HDA coexist [19,20] as mentioned in the manuscript, but HDA was not observed in this study. The authors should describe the difference from the previous study.  
+
+At the end of the section on crystallization of ice VII in the stability domain of ice VI, we explicitly mention that ice VII transforming into ice VI could evolve to a mixture of ice VI and HDA. We have no experimental evidence for a more ascertain statement.  
+
+Minor points  
+
+1. The precise wavelength should be described in Method section. I found the \(x\) -ray energy is \(19 \mathrm{keV}\) , but the number of significant figures is only two in this case.  
+2. The reflection indices should not be in parenthesis. Index in parenthesis, (hkl) means Miller index that is used for corresponding Miller plane in real space. But reflection index, hkl, in reciprocal space should not have any parenthesis (see International Tables for crystallography, for this criterion).  
+
+The manuscript was modified accordingly.  
+
+3. Effective digit and/or the error should be considered throughout the whole manuscript, in particular, in Table I.  
+
+This comment was taken into account throughout the whole manuscript.
+
+<--- Page Split --->
+
+## REVIEWER COMMENTS  
+
+## Reviewer #1 (Remarks to the Author):  
+
+The authors have adequately responded to my comments on their first submission. I'm happy to recommend publication.  
+
+## Reviewer #2 (Remarks to the Author):  
+
+This manuscript has clarified the occurrence of crystallization into ice VII occurs in between19  
+
+1.6 GPa and 2.0 GPa, that is in the stability field of ice VI. This was achieved by employing time-resolved X-ray diffraction and a Diamond-Annil-Cell, with pressure rise times from 0.1 ms to 100 ms.  
+
+The authors have added appropriate references as suggested by the reviewers. The authors have also added appropriate clarifying sentences in response to reviewer's remarks and questions.  
+
+An important exception to this is an incorrect comment made by the authors regarding an important reference to an article by cited by a reviewer. The authors state in their rebuttal comments "We now have cited the Yamamoto reference in our supplementary information" This is not seen in the revised manuscript and supplementary information copy submitted by the authors.  
+
+In summary, this revised manuscript describes in more detail than previously published studies the results of a re- examination of the phase transition of water to ice VII. It is not clear however that providing additional details of a known phase transition qualifies for acceptance in Nature.  
+
+## Reviewer #3 (Remarks to the Author):  
+
+Please find the attached comments.  
+
+[editorial note: please see the next page(s) for reviewer #3's comments.]
+
+<--- Page Split --->
+
+Comments on the manuscript entitled 'Metastable water at several compression rates and its freezing kinetics into ice VII" by Pepin et al.,  
+
+I noticed other two reviewers appreciated the importance of this study, which is in itself pleasing as a researcher in ice study. Although I understand most of authors rebuttal comments, authors replies to my comments raised the following further concerns. I point out my concerns as replies (bold) following to the rebuttal comments from the authors (italic).  
+
+1.  
+
+Regarding the solidification of ice VII in the stability domain of ice VI, we thank reviewer#3 for providing this Jp. J. Appl. Phys. reference. However, in it, the reported observation of ice VII in the stability field of ice VI is not exactly what is reported in our study. The nucleation of ice VII was obtained at the metastable extension of the melting line of ice VII in the stability field of ice VI by heating the sample so that the overcompression of water that can be obtained under quasi- static compression could reach the metastable melting line of ice VII. Nucleation of ice VII is stated then to occasionally form and a facetted single crystal of ice VII can bestabilized in equilibrium with the liquid. However, ice VII is kept metastable only if facetted in equilibrium with the liquid. That is exactly what we observed and used in the experiment reported in our supplementary material to precisely extend the ice VII melting line in the stability domain of ice VI. We now have cited the reference in our supplementary information.  
+
+However, out of this liquid- solid phase equilibrium, by slightly increasing pressure ice VII transforms into ice VI. We are reported a different story here. Under dynamical compression, metastable water always freeze into ice VII in the metastability domain of ice VII and then ice VII transforms to ice VI over a ms time scale and a mixture ice VI - ice VII then exists, which should eventually evolve to an amorphous - ice VI mixture. This resolves all previous contradictions of the literature between reports of amorphous HAD (could be HDA?), HAD- ice VI mixture or ice VII.  
+
+and the reply to the last comment.  
+
+- Previous studies using d-DAC reported that ice VI and HDA coexist [19,20] as mentioned in the manuscript, but HDA was not observed in this study. The authors should describe the difference from the previous study. At the end of the section on crystallization of ice VII in the stability domain of ice VI, we explicitly mention that ice VII transforming into ice VI could evolve to a mixture of ice VI and HDA. We have no experimental
+
+<--- Page Split --->
+
+I agree the dynamic compression may be somehow different to the static compression, but one point I would indicate is that the metastable ice VII made by the static compression in ice VI stability region also eventually convert to ice VI, since it is metastable. The difference to the static compression would be the degree of over- compression (AG between liquid and ice VII). Facet morphology found in ice VII by the static compression could also be found in that made by the dynamic compression in sub- micrometer scale, which could be just the difference of size. In that sense, the phenomenological difference between static and dynamic compression may be still unclear.  
+
+In this study, the HDA- ice VI mixture was not observed, and may not be explained the clear reason why HDA was observed in the previous study as stated in the reply to the last comment. It is briefly mentioned in the line 190- 197, as 'This difference could be explained by the fact that at 1.8 GPa, ice VII is approaching its thermodynamic stability, allowing for the stabilization of a mixture of ice VI and ice VII, whereas at 1.7 GPa, ice VII is gradually transforming, leading to the stabilization of a mixture of ice VI and amorphous ice.' But I could not satisfy this description resolves all previous contractions. The present results would show even another contradiction in the sense that they could not observe amorphous state.  
+
+2.  
+
+We thus could show that the heating of the sample is negligible. That is illustrated in the figure below showing that, no discernible heating was detected within the measurement's sensitivity, approximately \(\pm 3 \text{K}\) . Therefore, we can confidently assert that there is no heating occurring. We have now included this quantification of the absence of heating in the supplementary material.
+
+<--- Page Split --->
+![PLACEHOLDER_20_0]
+
+<center>Dynamic compression of liquid water at a compression rate (10 GPa/ms). The sample pressure over time is measured using simultaneously the S(Text(Si6)_[4]/text(O)_[7]/text(Sm)^(2+)/S and the ruby fluorescence </center>  
+
+From the above figure, I could see significant increases in temperatures, exceeding 315 K (42 °C), it does not look like \(\pm 3\) K. Plus, the plot in Fig. 9 shows the temperature \(\sim 21\) °C (294 K), it contradicts with the above figure. If the temperature changes over 10 K, the story might change since the experimental conditions are close to the melting line.  
+
+![PLACEHOLDER_20_1]
+  
+
+Finally one minor question for the above figure, what is the blue atom in ice VI structure?
+
+<--- Page Split --->
+
+## Detailed answer to reviewers' comments.  
+
+## Reviewer #1:  
+
+The authors have adequately responded to my comments on their first submission. I'm happy to recommend publication.  
+
+We thank reviewer #1 for his recommendation.  
+
+## Reviewer #2:  
+
+This manuscript has clarified the occurrence of crystallization into ice VII occurs in between 1.6 GPa and 2.0 GPa, that is in the stability field of ice VI. This was achieved by employing time- resolved X- ray diffraction and a Diamond- Anvil- Cell, with pressure rise times from 0.1 ms to 100 ms.  
+
+The authors have added appropriate references as suggested by the reviewers. The authors have also added appropriate clarifying sentences in response to reviewer's remarks and questions. An important exception to this is an incorrect comment made by the authors regarding an important reference to an article by cited by a reviewer. The authors state in their rebuttal comments "We now have cited the Yamamoto reference in our supplementary information" This is not seen in the revised manuscript and supplementary information copy submitted by the authors.  
+
+In summary, this revised manuscript describes in more detail than previously published studies the results of a re- examination of the phase transition of water to ice VII. It is not clear however that providing additional details of a known phase transition qualifies for acceptance in Nature.  
+
+We thank reviewer #2 for his feedback on our revised manuscript. We acknowledge the oversight regarding the the Yamamoto reference and have now included it in the revised supplementary information.  
+
+However, we are surprised by the change in Reviewer #2's appreciation of our manuscript. Initially, the reviewer suggested acceptance after minor comments were addressed, but he/she is now questioning the suitability of our work for publication in Nature. This change in stance is unexpected and lacks clarification.  
+
+## Reviewer #3:  
+
+I noticed other two reviewers appreciated the importance of this study, which is in itself pleasing as a researcher in ice study. Although I understand most of authors rebuttal comments, authors replies to my comments raised the following further concerns  
+
+I agree the dynamic compression may be somehow different to the static compression, but one point I would indicate is that the metastable ice VII made by the static compression in ice VI
+
+<--- Page Split --->
+
+stability region also eventually convert to ice VI, since it is metastable. The difference to the static compression would be the degree of over- compression (AG between liquid and ice VII). Facet morphology found in ice VII by the static compression could also be found in that made by the dynamic compression in sub- micrometer scale, which could be just the difference of size. In that sense, the phenomenological difference between static and dynamic compression may be still unclear. In this study, the HDA- ice VI mixture was not observed, and may not be explained the clear reason why HDA was observed in the previous study as stated in the reply to the last comment. It is briefly mentioned in the line 190- 197, as 'This difference could be explained by the fact that at 1.8 GPa, ice VII is approaching its thermodynamic stability, allowing for the stabilization of a mixture of ice VI and ice VII, whereas at 1.7 GPa, ice VII is gradually transforming, leading to the stabilization of a mixture of ice VI and amorphous ice.' But I could not satisfy this description resolves all previous contractions. The present results would show even another contradiction in the sense that they could not observe amorphous state.  
+
+We respectfully disagree with reviewer #3. There is a phenomenological difference between static and dynamic compressions concerning the observation of stable ice VII in the stability field of ice VI. In the static experiment, ice VII was brought into the stability field of ice VI by following the metastable melting line of ice VII down to 300K. If the melting equilibrium is lost and complete solidification occurs, ice VII transforms to ice VI. In dynamical compression, the over- compression pressure of metastable water is increasing with the compression rate. For compression rates in between 0.15 GPa/ms sans 0.9 GPa/ms, the freezing of metastable water falls in the stability domain of ice VI and in the metastability domain of ice VII, as determined by the extension of the melting line of ice VII. Our study shows that metastable water under such conditions always freezes in ice VII and then transforms into ice VI over a few milliseconds. Using time resolved XRD we could thus show a preferred nucleation mechanism into ice VII which was not shown by static measurements.  
+
+Regarding the formation of amorphous ice, HDA, as reported by previous studies based on Raman measurements, we are unable to observe it because it is a too weak signal by X- ray diffraction. It should also be noted that a mixture of fine powder of ice VI and ice VII would have a Raman signal that looks like the one of HDA ice. The important points of our study are the preferred nucleation of water into ice VII and then the time- scale of the transformation into ice VI. Whether there is complete transformation into ice VI, a mixture of ice VI and ice VII, a mixture of ice VI and HDA could depend of the final pressure of the rapid compression. We have modified the manuscript to make this point clearer. The sentence now reads: "This difference could be explained by the fact that at 1.8 GPa, ice VII is approaching its thermodynamic stability, allowing for the stabilization of a mixture of ice VI and ice VII (and a possible amorphous state, which could not be observed in our case due to its very low diffraction power), whereas at 1.7 GPa, ice VII has been reported to gradually transform in a mixture of ice VI and amorphous ice."  
+
+From the above figure, I could see significant increases in temperatures, exceeding 315 K (42 °C), it does not look like \(\pm 3\) K. Plus, the plot in Fig. 9 shows the temperature \(\sim 21\) °C (294 K), it contradicts with the above figure. If the temperature changes over 10 K, the story might change since the experimental conditions are close to the melting line.
+
+<--- Page Split --->
+![PLACEHOLDER_23_0]
+  
+
+We apologize for any confusion caused by the previous error bars in our temperature measurements. The \(\pm 3 \mathrm{K}\) error bar was taken from the reference article based on the luminescence gauges, but we have since re- estimated the error bars to be \(\pm 10 \mathrm{K}\) .  
+
+We would like to point out that the temperature relation in our study was obtained in non- hydrostatic pressure conditions, whereas in ref. 29, the error bar estimation was obtained in a hydrostatic pressure medium. This impacts the fluorescence of the ruby by broadening the peaks and making them asymmetrical, which slightly skews the pressure measurements. This is evidenced in the bottom graph of the figure provided. If a temperature increase were to take place during the fast compression, the temperature should be an increasing function of time throughout the entire compression sequence. However, this is not the case here, as we observed the temperature oscillating around the initial temperature of \(294 \mathrm{K}\) , within a \(\pm 10 \mathrm{K}\) error bar.  
+
+Furthermore, if we consider a temperature increase of \(10 \mathrm{K}\) for our point under 14.82 GPa/ms, the figure below shows that it corresponds to a pressure difference in the overcompression of metastable water of \(\sim 0.06 \mathrm{GPa}\) , which is within the error bars that we considered in our models of figures 5 and 6.  
+
+In conclusion, the temperature increase during compression is very small and if any, of the order of \(10 \mathrm{K}\) , does not affect the overall results and conclusions of our study.  
+
+We have slightly modified the manuscript to reflect this and the figure as shown above. The sentence now reads: "That is illustrated by the figure below showing that, no discernible heating was detected within the measurement's sensitivity, approximately \(\pm 10 \mathrm{K}\) ".
+
+<--- Page Split --->
+![PLACEHOLDER_24_0]
+  
+
+Finally one minor question for the above figure, what is the blue atom in ice VI structure?  
+
+This is just a visual representation of the oxygen atoms occupying two different positions (Wycoff site 2b in blue and 8g in red).
+
+<--- Page Split --->
+
+## REVIEWERS' COMMENTS  
+
+Reviewer #3 (Remarks to the Author): Please find the attached comments.  
+
+[editorial note: please see the next page(s) for reviewer #3's comments.]
+
+<--- Page Split --->
+
+1. I acknowledge the author's comments on my previous concerns regarding the phenomenological difference between static and dynamic compression. First, I would indicate one fact that ice VII could also nucleate into the stability field of ice VI as written by Yamamoto et al., described as  
+
+"...it was discovered in the present experiment that ice VII occasionally formed directly from water at point B in Fig. 1 without the formation of ice VI at point A." (see below)  
+
+[figure redacted]  
+
+Therefore, the author's indication in their rebuttal comments,  
+
+"In the static experiment, ice VII was brought into the stability field of ice VI by following the metastable melting line of ice VII down to 300K. If the melting equilibrium is lost and complete solidification occurs, ice VII transforms to ice VI."  
+
+is misunderstanding.  
+
+But, if in the dynamic compression ice VII "always" preferably nucleates rather than ice VI, this is the phenomenological difference since in the static compression ice VII could only "occasionally" nucleate. Authors could emphasize this point with the reference to Yamamoto et al.
+
+<--- Page Split --->
+
+2. As for the description of (non)appearance of HDA, the added sentence "(and a possible amorphous state, which could not be observed in our case due to its very low diffraction power)" is a bit inaccurate since the diffraction power ("scattering power" may be more appropriate term in the case of amorphous, though "diffraction power" is also correct) itself is comparable to the crystalline material if all integrated through whole reciprocal space. I suggest this sentence would be "...in our case since the scattering from amorphous may be hidden in the Bragg peaks even it exists" or so. In any case, the last sentence of this paragraph, "These results offer a clarification to the apparent contradiction of the previous findings." seems a bit overstatement to me.  
+
+3. I consider, repeatedly saying, that the difference to static compression would be the degree of overcompression. The over-compression as large as a few GPa could not be achieved by the static compression. The strong point of this study would be the quantitative analysis of the nucleation time as a function of the over-compression. This is the reason why I repeatedly indicate the quantitative considerations for the temperature effect since I suppose the \(10\mathrm{K}\) difference is huge for water molecules at a few GPa regions. I suggest some comments on the temperature effect, as written in the rebuttal comments, may be included in the main text as well.  
+
+4. Regarding the qualitative analysis, I pointed out in the first comment,  
+
+"Effective digit and/or the error should be considered throughout the whole manuscript, in particular, in Table I."  
+
+and authors replied,  
+
+"This comment was taken into account throughout the whole manuscript."  
+
+I did not carefully check in the second round (since I believed this reply), but I realized that Table I was not modified yet. For example, the rate of compression in the first raw is written as '870787.5 (GPa/ms)". I could not believe the effective digit could reach 7 digits, such very precise pressure control would not be possible. Similarly, in the caption of Figure 6, the parameter b is written like "b=245293534.93", which is also unrealistic effective digit considering the fitting result shown in Figure 6. Such representations without care of effective digits led readers to the question of the whole results of this study. Authors should carefully check again this point.
+
+<--- Page Split --->
+
+## Detailed answer to reviewers' comments.  
+
+1- I acknowledge the author's comments on my previous concerns regarding the phenomenological difference between static and dynamic compression. First, I would like to point out that ice VII could also nucleate within the stability field of ice VI, as written by Yamamoto et al., described as:  
+
+"...it was discovered in the present experiment that ice VII occasionally formed directly from water at point B in Fig. I without the formation of ice VI at point A." (see below) Therefore, the author's statement in their rebuttal comments, "In the static experiment, ice VII was brought into the stability field of ice VI by following the metastable melting line of ice VII down to 300K. If the melting equilibrium is lost and complete solidification occurs, ice VII transforms to ice VI." is a misunderstanding. However, if under dynamic compression ice VII "always" nucleates preferentially rather than ice VI, this would be the phenomenological difference, since under static compression ice VII could only "occasionally" nucleate. The authors could emphasize this point with reference to Yamamoto et al.  
+
+Following the reviewer's advice, we have added a sentence that reads: "It was also shown that ice VII could occasionally nucleate in the stability domain of ice VI above the metastable melting line [41]".  
+
+2- Regarding the description of the (non)appearance of HDA, the added sentence "(and a possible amorphous state, which could not be observed in our case due to its very low diffraction power)" is somewhat inaccurate, as the diffraction power ("scattering power" might be a more appropriate term in the case of amorphous materials, although "diffraction power" is also correct) itself is comparable to that of crystalline material if integrated over the whole reciprocal space. I suggest rephrasing this sentence as "...in our case since the scattering from amorphous material may be hidden in the Bragg peaks even if it exists," or something similar. In any case, the last sentence of this paragraph, "These results offer a clarification to the apparent contradiction of the previous findings," seems a bit of an overstatement to me.  
+
+We have modified the main text accordingly.  
+
+3- I believe, and have repeatedly stated, that the difference with static compression lies in the degree of over-compression. Over-compression as large as a few GPa could not be achieved by static compression. The strength of this study lies in the quantitative analysis of nucleation time as a function of over-compression. This is why I repeatedly emphasize the importance of quantitative considerations regarding the temperature effect, as I believe the 10 K difference is significant for water molecules under a few GPa. I suggest that some comments on the temperature effect, as written in the rebuttal comments, be included in the main text as well.  
+
+We have added a statement in the supplementary information. It reads: "We would like to point out that the temperature relation in our study was obtained in non- hydrostatic pressure conditions, whereas in Ref.\cite{Datchi97}, the error bar estimation was obtained in a hydrostatic pressure medium. This impacts the fluorescence of the ruby by broadening the peaks and making them asymmetrical, which slightly skews the pressure measurements. This is evidenced in the bottom graph of Supplementary Figure \ref{fig:figureS4}. If a temperature increase were to take place during the fast compression, the temperature should be an
+
+<--- Page Split --->
+
+increasing function of time throughout the entire compression sequence. However, this is not the case here, as we observed the temperature oscillating around the initial temperature of 294 K, within a \(\pm 10 \mathrm{K}\) error bar.  
+
+Furthermore, if we consider a temperature increase of \(10 \mathrm{K}\) for the point at \(14.82 \mathrm{GPa / ms}\) , it would correspond to a pressure difference in the overcompression of metastable water of \(- 0.06 \mathrm{GPa}\) , which is within the error bars that we considered in our models of figures 5 and 6. ”  
+
+4- Regarding the qualitative analysis, I pointed out in my first comment, "Effective digits and/or errors should be considered throughout the entire manuscript, particularly in Table I," and the authors replied, "This comment was taken into account throughout the entire manuscript." I did not carefully check this during the second round (as I trusted their reply), but I later realized that Table I had not yet been modified. For example, the rate of compression in the first row is written as '870787.5 (GPa/ms).' I cannot believe the effective digits could reach seven digits; such precise pressure control would not be possible. Similarly, in the caption of Figure 6, the parameter \(b\) is written as " \(b = 245293534.93\) ," which is also an unrealistic number of significant digits considering the fitting result shown in Figure 6. Such representations without careful attention to significant digits lead readers to question the overall results of this study. The authors should carefully check this point again.  
+
+We have modified the main text accordingly, carefully checking for effective digits. However, some number have not been modified because they are derived from data from references cited in our manuscript.
+
+<--- Page Split --->

peer_reviews/127b4a1df348ee954b5fefb88eeed0676c975caa5b61439cfcd7368eac65b846/supplementary_0_Peer Review File/supplementary_0_Peer Review File_det.mmd

@@ -0,0 +1,619 @@
+<|ref|>title<|/ref|><|det|>[[61, 40, 506, 90]]<|/det|>
+# nature portfolio  
+
+<|ref|>text<|/ref|><|det|>[[70, 110, 361, 139]]<|/det|>
+Peer Review File  
+
+<|ref|>text<|/ref|><|det|>[[70, 155, 934, 210]]<|/det|>
+Metastable water at several compression rates and its freezing kinetics into ice VII  
+
+<|ref|>image<|/ref|><|det|>[[57, 732, 240, 784]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[250, 732, 911, 784]]<|/det|>
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[118, 84, 843, 129]]<|/det|>
+Editorial note: Parts of this Peer Review File have been redacted as indicated to remove third- party material where no permission to publish could be obtained.  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 182, 315, 199]]<|/det|>
+## REVIEWER COMMENTS  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 234, 437, 252]]<|/det|>
+## Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[117, 259, 877, 383]]<|/det|>
+This work shows in situ, time resolved x- ray diffraction data on the H2O system subjected to a wide range of compression rates. The diffraction data show real- time structural response of the system and allow unambiguous identification of solid phases of ice. I think the presented data provide interesting and important insight into the freezing dynamics and help to resolve discrepancies from other measurements.  
+
+<|ref|>text<|/ref|><|det|>[[118, 415, 844, 460]]<|/det|>
+I had several thoughts on how the manuscript could be made clearer but, beyond these, would like to recommend publication.  
+
+<|ref|>text<|/ref|><|det|>[[115, 493, 875, 588]]<|/det|>
+1. Page 2 introduction: "having the simplest ice structure with..." I think the work "simplest" is subjective. Since the structure of ice VII is highly disordered, one could say an ordered phase, such as ice VIII is structurally simpler. More appropriate would be a specific statement like "simplest crystallographic unit cell".  
+
+<|ref|>text<|/ref|><|det|>[[117, 597, 877, 642]]<|/det|>
+2. Page 2 introduction: "is a topic of current focus" when stating this, I think citations should be provided.  
+
+<|ref|>text<|/ref|><|det|>[[117, 651, 833, 722]]<|/det|>
+3. Page 5 instrumentation: It is mentioned that the pressure is measured via multiple approaches. Hence, how does the reader know which is refered to when pressures are quoted throughout the manuscript. E.g. what is shown in Figure 1?  
+
+<|ref|>text<|/ref|><|det|>[[117, 730, 860, 774]]<|/det|>
+4. Page 5 pressure response of...compression ramps": "the convolution with" I'm not sure this effect is a convolution in the mathematical sense.  
+
+<|ref|>text<|/ref|><|det|>[[117, 782, 874, 878]]<|/det|>
+5. Page 5 pressure response of...compression ramps": "1.56 GPa" No quantified uncertainty is given for this number and, on the basis of Figure 2 where error bar are shown to be ±0.05 this number is stated too precisely. Assuming the error bars indicate the magnitude of a standard deviation, quoting the pressure to one decimal place is sufficient.  
+
+<|ref|>text<|/ref|><|det|>[[115, 886, 872, 905]]<|/det|>
+6. Page 8 Freezing pressure of...compression rate: "0.056, which matches...0.069" Use of the
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[117, 84, 820, 128]]<|/det|>
+word 'matches' requires some information on experimental uncertainty, which is not provided.  
+
+<|ref|>text<|/ref|><|det|>[[117, 137, 860, 207]]<|/det|>
+7. Page 8 Freezing pressure of..compression rate: "When metastable water freeze under a compression rate of 110 GPa/ms, it corresponds to a cooling rate of \(10^{\wedge}7 \text{K / s}\) ". I have no idea how this equivalence was derived, can the authors explain their argument?  
+
+<|ref|>text<|/ref|><|det|>[[117, 215, 861, 310]]<|/det|>
+8. Figure 2: I was confused by the meaning of the coloured horizontal bands indicating phase regions. Are these taken from other work? Or if they are they guides to the eye to indicate behaviour observed in this study, I found it confusing that they are horizontal and not vertical as each phase is only present for bounded periods of time.  
+
+<|ref|>text<|/ref|><|det|>[[117, 318, 876, 386]]<|/det|>
+9. Figure 4: How can pressure be inferred from this plot? Is it possible to show in some way? 10. Figure 4: Out of curiosity, is some interpolation or smoothing applied to the 2d 2theta vs time data?  
+
+<|ref|>text<|/ref|><|det|>[[117, 396, 872, 546]]<|/det|>
+11. Figure 4: "A mixture of ice VI and ice VII is still observed" I wondered if it is possible by Rietveld analysis (likely with a highly constrained model) to determine the phase fraction of VI and VII? From the 2d image, it indeed looks like the intensity of the ice VII 110 decreases as the ice VI forms. If it could be quantitatively shown that this change is due to direct transformation of the volume of ice VII in the x-ray beam, that would strengthen the interpretation given.  
+
+<|ref|>sub_title<|/ref|><|det|>[[120, 606, 438, 623]]<|/det|>
+## Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[118, 632, 870, 728]]<|/det|>
+This manuscript describes the effects of rapid compression of liquid water to its crystalline phases using dynamic- piezo- Diamond- Anvil- Cell. The main goal of this study was to provide experimental details and perhaps a physical understanding the structural changes of water undergoing rapid temperature and/or pressure variation.  
+
+<|ref|>text<|/ref|><|det|>[[117, 761, 880, 910]]<|/det|>
+Although there have been many studies of phase transitions in both water and ice, this study focused on one particular phase transition where there have been different previously reported results and interpretations. This is a transition between metastable liquid water and particular dense crystalline phases. The authors of this study employed additional an experimental technique of time- resolved X- ray diffraction with a dynamic- piezo- Diamond- Anvil- Cell. The employment of this technique provided sufficient new data to resolve the
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[118, 85, 446, 101]]<|/det|>
+contradictions that are in the literature.  
+
+<|ref|>text<|/ref|><|det|>[[115, 135, 880, 707]]<|/det|>
+This is a very well written manuscript and I therefore have only a small list of comments that should be addressed in more detail.1) On page 3 the authors indicate that they are investigating whether ice VII is still the freezing state of metastable water?? Under what, P conditions are they referring to?2) On page 5 the authors give a value of 1.56 GPa? How accurate is this number? T, P effects?3) On page 6: It was stated and suggested that crystallization occurs into ice VII, although without time-resolved XRD measurements this proposition remains to be proven. Ref?4) Pg. 6: excellent agreement between the different pressure determinations. Can the authors be clearer on this point? Perhaps by simply referring to the figures.5) On Pg. 7, it is stated that there was a e previous domain of investigation in d- DAC experiments. A reference. is needed here.6) without time-resolved XRD measurements this proposition remains to be proven. A reference should be made here.7) On page. 9 the authors state "Since water and ice are quite incompressible" Here a comparison with other triatomic molecules would strengthen this statement.8) It is important to emphasize that the model employed to fit the data is a classical nucleation theory description of the experimental results. It fits the data very well and provides a very good physical picture and interpretation of the experimental results but this depends on parameters used for fitting the data so small changes in the parameters employed could have a significant effect on the model used. This could be mentioned by the authors.  
+
+<|ref|>text<|/ref|><|det|>[[117, 787, 878, 911]]<|/det|>
+This report follows up on the previous domain of investigation in d- DAC experiments focusing on the solidification of metastable water. This is the key goal reached in this study. The authors clearly state that support the hypothesis that metastable water, within the pressure range of 1.5 GPa to 2.1 GPa, nucleates into ice VII first, despite this range being the stability field of ice VI. results offer a clarification to the apparent contradiction of the
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[118, 85, 265, 101]]<|/det|>
+previous findings.  
+
+<|ref|>text<|/ref|><|det|>[[117, 110, 880, 234]]<|/det|>
+previous findings.In summary, this manuscript clearly addresses and presents a solution to the phase transition discrepancies in the literature. A key improvement in the measurement technique was time- resolved X- ray diffraction. This technique is very well described in the manuscript All figures were clearly presented and definitely provided added clarity to the report. Acceptance is therefore suggested after the minor comments listed above are addressed.  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 292, 438, 310]]<|/det|>
+## Reviewer #3 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 317, 875, 600]]<|/det|>
+The authors investigated the interesting behaviour of water ice under dynamic compression with various compression rates using d- DAC by time- resolved x- ray diffraction method and found that ice VII crystallise under the conditions where ice VI should be stable. They also interpret the excess pressure and growth time during the crystallization of ice VII using a phenomenological model based on the classical nucleation theory, succeeding in a comprehensive explanation for the growth rate of ice VII grown under various compression rates. The investigated technique, x- ray diffraction with ruby fluorescence observations at micro- to milli- second order, is somehow remarkable, and the overall achievements of this study would be worth recognising for the ice community. Thus, I do not doubt that this manuscript should be eventually published in at least some specific journal, but I am not sure whether this manuscript will be published by Nature Communication.  
+
+<|ref|>text<|/ref|><|det|>[[115, 606, 877, 912]]<|/det|>
+Two key points in assessing the value of this study would be 1) how new the observed results are and 2) how much their interpretation will influence other related studies. First, concerning the observed results, it should be stated that no truly new phenomena have been reported in this study. It has long been known among ice researchers that ice VII nucleates in the stable region of ice VI even under static compression (e.g. K. Yamamoto, Jpn J Appl Phys, 19, 1841, 1980, doi: 10.1143/JJAP.19.1841). A recent study using d- DAC has also reported ice VII nucleation in the stability region of ice VI [18]. Therefore, what is new in this manuscript are the results of quantitative experiments on how the growth rate of ice VII changes when the compression rate is varied. The experimental results in this study are novel in the sense that they are quantitative, but the observed phenomena themselves are not particularly novel. The manuscript also does not elaborate on how the quantitative growth rates obtained might affect other related studies. It is therefore difficult to assess, at
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[117, 85, 565, 101]]<|/det|>
+least at the moment, how important these results are.  
+
+<|ref|>text<|/ref|><|det|>[[117, 110, 861, 181]]<|/det|>
+Even if the importance of this manuscript is recognised by other reviewers and considered for publication in Nat. Commun., the following concerns remain and should be fully considered before potential publication.  
+
+<|ref|>text<|/ref|><|det|>[[116, 188, 866, 365]]<|/det|>
+- The authors assume that the sample temperature should not be increased by the rapid compression, but it should be more carefully evaluated and explained in the manuscript. I also believe that temperature might be quickly stable at room temperature owing to the high thermal conductivity of the diamond. However, if the sample size is enough large, this might not be the case since the thermal conductivity of ice VII would not be such high, so the temperature could temporally increase in a very short time. This issue should be quantitatively assessed, for example, by FEM simulation.  
+
+<|ref|>text<|/ref|><|det|>[[115, 396, 879, 677]]<|/det|>
+- In the adopted phenomenological model for the nucleation rate, authors ignore the effect of crystal growth, but only take the nucleation rate into account. However, I recognized that the grown ice VII may be somehow coarse crystalline size because the observed diffraction patterns are not like smooth lines but spotty. In particular, I found 111 spots of ice VII, which should be very very weak (generally invisible) if the specimen is fine powder, showing the specimen should be somehow coarse aggregates of many single crystals. The actual rate of ice VII growth may be expressed as the combination of nucleation and growth rates. In fact, the results obtained in this study could be interpreted by different phenomenological models, such as the simple JMAK model. The point would be 'which model is the best' to describe the observed phenomena. Even if one model fits well, this does not necessarily mean that this is the only correct model.  
+
+<|ref|>text<|/ref|><|det|>[[117, 684, 877, 782]]<|/det|>
+Plus, I am also concerned about the lack of temperature terms in the adopted model. I suppose the temperature term should appear since the classical nucleation theory generally has it. Authors should discuss this point (why temperature term could be ignored), if they stick to this model only.  
+
+<|ref|>text<|/ref|><|det|>[[117, 814, 872, 885]]<|/det|>
+- Previous studies using d-DAC reported that ice VI and HDA coexist [19,20] as mentioned in the manuscript, but HDA was not observed in this study. The authors should describe the difference from the previous study.
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[118, 86, 227, 101]]<|/det|>
+## Minor points  
+
+<|ref|>text<|/ref|><|det|>[[115, 110, 870, 315]]<|/det|>
+1. The precise wavelength should be described in Method section. I found the x-ray energy is 19 keV, but the number of significant figures is only two in this case.  
+2. The reflection indices should not be in parenthesis. Index in parenthesis, (hkl) means Miller index that is used for corresponding Miller plane in real space. But reflection index, hkl, in reciprocal space should not have any parenthesis (see International Tables for crystallography, for this criterion).  
+3. Effective digit and/or the error should be considered throughout the whole manuscript, in particular, in Table I
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[115, 84, 528, 102]]<|/det|>
+## Detailed answer to reviewers' comments.  
+
+<|ref|>sub_title<|/ref|><|det|>[[115, 121, 222, 138]]<|/det|>
+## Reviewer #1  
+
+<|ref|>text<|/ref|><|det|>[[115, 155, 880, 244]]<|/det|>
+This work shows in situ, time resolved \(x\) - ray diffraction data on the H2O system subjected to a wide range of compression rates. The diffraction data show real- time structural response of the system and allow unambiguous identification of solid phases of ice. I think the presented data provide interesting and important insight into the freezing dynamics and help to resolve discrepancies from other measurements.  
+
+<|ref|>text<|/ref|><|det|>[[115, 261, 880, 296]]<|/det|>
+I had several thoughts on how the manuscript could be made clearer but, beyond these, would like to recommend publication.  
+
+<|ref|>text<|/ref|><|det|>[[115, 314, 880, 385]]<|/det|>
+1. Page 2 introduction: "having the simplest ice structure with..." I think the work "simplest" is subjective. Since the structure of ice VII is highly disordered, one could say an ordered phase, such as ice VIII is structurally simpler. More appropriate would be a specific statement like "simplest crystallographic unit cell".  
+
+<|ref|>text<|/ref|><|det|>[[117, 394, 459, 412]]<|/det|>
+The sentence was modified accordingly.  
+
+<|ref|>text<|/ref|><|det|>[[115, 439, 880, 474]]<|/det|>
+2. Page 2 introduction: "is a topic of current focus" when stating this, I think citations should be provided.  
+
+<|ref|>text<|/ref|><|det|>[[117, 484, 333, 501]]<|/det|>
+We have added citations.  
+
+<|ref|>text<|/ref|><|det|>[[115, 529, 822, 582]]<|/det|>
+3. Page 5 instrumentation: It is mentioned that the pressure is measured via multiple approaches. Hence, how does the reader know which is referred to when pressures are quoted throughout the manuscript. E.g. what is shown in Figure 1?  
+
+<|ref|>text<|/ref|><|det|>[[115, 599, 870, 652]]<|/det|>
+As mentioned in the manuscript, the imaging and X- ray diffraction (XRD) processes could not be conducted simultaneously. The pressure was measured using luminescence during imaging and by determining the volume of copper during XRD.  
+
+<|ref|>text<|/ref|><|det|>[[115, 666, 833, 717]]<|/det|>
+An essential aspect to highlight, as illustrated in Figure 3, is the high degree of repeatability and reproducibility of the compression ramps using the different diagnostic techniques. To enhance clarity, we have included the following sentence:  
+
+<|ref|>text<|/ref|><|det|>[[115, 731, 868, 782]]<|/det|>
+"The pressure was determined using a luminescence gauge during imaging and by utilizing XRD equation- of- state (EOS) data during XRD. The results from the different pressure measurement methods were found to be highly consistent with each other."  
+
+<|ref|>text<|/ref|><|det|>[[115, 815, 857, 850]]<|/det|>
+4. Page 5 pressure response of...compression ramps": "the convolution with" I'm not sure this effect is a convolution in the mathematical sense.  
+
+<|ref|>text<|/ref|><|det|>[[115, 860, 848, 913]]<|/det|>
+The sentence now reads: "However, if a phase transition occurs, the pressure rise is disrupted by the pressure drop associated to the negative volume discontinuity at the transition, resulting in an apparent negative compressibility."
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[116, 100, 877, 171]]<|/det|>
+5. Page 5 pressure response of...compression ramps": "1.56 GPa" No quantified uncertainty is given for this number and, on the basis of Figure 2 where error bar are shown to be \(\pm 0.05\) this number is stated too precisely. Assuming the error bars indicate the magnitude of a standard deviation, quoting the pressure to one decimal place is sufficient.  
+
+<|ref|>text<|/ref|><|det|>[[116, 181, 861, 199]]<|/det|>
+The pressure was modified to 1.6 GPa to take into account the \(\pm 0.05\) GPa uncertainty.  
+
+<|ref|>text<|/ref|><|det|>[[116, 226, 875, 278]]<|/det|>
+6. Page 8 Freezing pressure of...compression rate: "0.056, which matches...0.069" Use of the word 'matches' requires some information on experimental uncertainty, which is not provided.  
+
+<|ref|>text<|/ref|><|det|>[[116, 288, 860, 323]]<|/det|>
+The sentence now reads: "Remarkably, the exponent 'c' obtained from the fit is 0.056, which closely aligns with the value of 0.069 predicted by Myint's scaling law."  
+
+<|ref|>text<|/ref|><|det|>[[116, 351, 846, 402]]<|/det|>
+7. Page 8 Freezing pressure of compression rate: "When metastable water freeze under a compression rate of 110 GPa/ms, it corresponds to a cooling rate of \(10^{7} \mathrm{~K} / \mathrm{s}\) ". I have no idea how this equivalence was derived, can the authors explain their argument.  
+
+<|ref|>text<|/ref|><|det|>[[115, 413, 880, 519]]<|/det|>
+At a compression rate of 110 GPa/ms at 300 K, the freezing process occurs at 2.9 GPa after an over-compression duration of 18 us in the metastable water state. To achieve the same state through a 18 us cooling at 2.9 GPa from the liquid state, considering the melting point of water at 2.9 GPa is 423 K (Datchi et al., PRB 61, 6535 (2000)), the required cooling rate would have to be \(7 \times 10^{6} \mathrm{~K} / \mathrm{s}\) , approximately \(10^{7} \mathrm{~K} / \mathrm{s}\) . This calculation has been included in the supplementary material.  
+
+<|ref|>text<|/ref|><|det|>[[115, 547, 877, 616]]<|/det|>
+8. Figure 2: I was confused by the meaning of the coloured horizontal bands indicating phase regions. Are these taken from other work? Or if they are they guides to the eye to indicate behaviour observed in this study, I found it confusing that they are horizontal and not vertical as each phase is only present for bounded periods of time.  
+
+<|ref|>text<|/ref|><|det|>[[115, 634, 877, 700]]<|/det|>
+The colored regions in the figure represent the (meta)stability pressure domains at 300 K for the various phases. These domains are inferred from the static phase diagram and the newly measured metastable melting of liquid- ice VII in this study. They are included as a visual aid to guide the interpretation of the data.  
+
+<|ref|>text<|/ref|><|det|>[[115, 733, 872, 750]]<|/det|>
+9. Figure 4: How can pressure be inferred from this plot? Is it possible to show in some way?  
+
+<|ref|>text<|/ref|><|det|>[[115, 760, 875, 830]]<|/det|>
+Pressure is inferred from the measured volumes of Sn, ice VII and ice VI. A reference to the Sn EoS has been added and a clarification has been added in the text which now reads: "Here, the ice sample remaining as a mixture of ice VI and ice VII is observed at 1.8 GPa, as inferred from the measured volumes of Sn, ice VII and VI".  
+
+<|ref|>text<|/ref|><|det|>[[115, 859, 869, 892]]<|/det|>
+10. Figure 4: Out of curiosity, is some interpolation or smoothing applied to the 2d 2theta vs time data?
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[117, 83, 533, 100]]<|/det|>
+A gourdash shading is indeed applied to the data.  
+
+<|ref|>text<|/ref|><|det|>[[115, 126, 875, 234]]<|/det|>
+11. Figure 4: "A mixture of ice VI and ice VII is still observed" I wondered if it is possible by Rietveld analysis (likely with a highly constrained model) to determine the phase fraction of VI and VII? From the 2d image, it indeed looks like the intensity of the ice VII 110 decreases as the ice VI forms. If it could be quantitatively shown that this change is due to direct transformation of the volume of ice VII in the x-ray beam, that would strengthen the interpretation given.  
+
+<|ref|>text<|/ref|><|det|>[[116, 240, 849, 312]]<|/det|>
+Although the presence of a mixture of ice VI and ice VII is undeniable, conducting a Rietveld analysis to determine the exact phase fractions of ice VI and ice VII appears impracticable here. This is mainly due to the fact that ice VI and ice VII exhibit different textures and the detector angle-coverage is only partial.  
+
+<|ref|>sub_title<|/ref|><|det|>[[116, 343, 224, 360]]<|/det|>
+## Reviewer #2  
+
+<|ref|>text<|/ref|><|det|>[[116, 378, 870, 450]]<|/det|>
+This manuscript describes the effects of rapid compression of liquid water to its crystalline phases using dynamic- piezo- Diamond- Anvil- Cell. The main goal of this study was to provide experimental details and perhaps a physical understanding the structural changes of water undergoing rapid temperature and/or pressure variation.  
+
+<|ref|>text<|/ref|><|det|>[[115, 466, 872, 590]]<|/det|>
+Although there have been many studies of phase transitions in both water and ice, this study focused on one particular phase transition where there have been different previously reported results and interpretations. This is a transition between metastable liquid water and particular dense crystalline phases. The authors of this study employed additional an experimental technique of time- resolved X- ray diffraction with a dynamic- piezo- Diamond- Anvil- Cell. The employment of this technique provided sufficient new data to resolve the contradictions that are in the literature.  
+
+<|ref|>text<|/ref|><|det|>[[115, 608, 868, 642]]<|/det|>
+This is a very well written manuscript and I therefore have only a small list of comments that should be addressed in more detail.  
+
+<|ref|>text<|/ref|><|det|>[[115, 679, 820, 714]]<|/det|>
+1) On page 3 the authors indicate that they are investigating whether ice VII is still the freezing state of metastable water?? Under what, P conditions are they referring to?  
+
+<|ref|>text<|/ref|><|det|>[[115, 723, 864, 884]]<|/det|>
+The previous speculations about the transformation of water into ice VII during shock compression experiments have been uncertain due to the lack of X- ray diffraction diagnostic. In this study, we aim to clarify this phenomenon by investigating the crystallization of metastable water under high compression rates within the stability field of ice VII. Specifically, we seek to determine whether the freezing state of metastable water under these conditions is indeed ice VII or another metastable structure. The text has been slightly modified to make it clearer: 'In particular, we are investigating whether the freezing state of metastable water at high compression rates above 2 GPa is indeed ice VII or a metastable structure....'
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 100, 820, 135]]<|/det|>
+2) On page 5 the authors give a value of1.56 GPa? How accurate is this number? T, P effects?  
+
+<|ref|>text<|/ref|><|det|>[[115, 144, 848, 181]]<|/det|>
+Referee #1 has also brought up a similar point. To address this, we have adjusted the pressure to 1.6 GPa, which accounts for the uncertainty of \(\pm 0.05\) GPa.  
+
+<|ref|>text<|/ref|><|det|>[[115, 207, 853, 243]]<|/det|>
+3) On page 6: It was stated and suggested that crystallization occurs into ice VII, although without time-resolved XRD measurements this proposition remains to be proven. Ref?  
+
+<|ref|>text<|/ref|><|det|>[[115, 252, 870, 305]]<|/det|>
+The relevant references (14- 17) have already been cited a few lines earlier. To avoid redundancy and maintain the flow of the text, we have chosen not to repeat them in this instance.  
+
+<|ref|>text<|/ref|><|det|>[[115, 333, 872, 368]]<|/det|>
+4) Pg. 6: excellent agreement between the different pressure determinations. Can the authors be clearer on this point? Perhaps by simply referring to the figures.  
+
+<|ref|>text<|/ref|><|det|>[[115, 377, 876, 430]]<|/det|>
+A reference to Figure 3 was added accordingly, showing that the pressure determination from the different gauges (Cu, Ice VII and SrB4O7 : Sm2+) are consistent with each other.  
+
+<|ref|>text<|/ref|><|det|>[[115, 457, 808, 492]]<|/det|>
+5) On Pg. 7, it is stated that there was a e previous domain of investigation in d-DAC experiments. A reference. is needed here.  
+
+<|ref|>text<|/ref|><|det|>[[115, 502, 594, 520]]<|/det|>
+The corresponding references (18- 20) have been added.  
+
+<|ref|>text<|/ref|><|det|>[[115, 547, 810, 582]]<|/det|>
+6) without time-resolved XRD measurements this proposition remains to be proven. A reference should be made here.  
+
+<|ref|>text<|/ref|><|det|>[[115, 592, 474, 610]]<|/det|>
+This point has been answered in point #3.  
+
+<|ref|>text<|/ref|><|det|>[[115, 636, 821, 672]]<|/det|>
+7) On page. 9 the authors state "Since water and ice are quite incompressible" Here a comparison with other triatomic molecules would strengthen this statement.  
+
+<|ref|>text<|/ref|><|det|>[[115, 681, 878, 824]]<|/det|>
+The bulk modulus of water ice is approximately 20 GPa, which is relatively high compared to other molecular systems, such as NH3 (4.2 GPa). However, the focus of this analysis is not on the incompressibility of the system, but rather on the assumption of constant compressibility. By using a first- order Birch equation of state (EoS), assuming constant compressibility, which is a reasonable approximation over the 10 GPa pressure range considered, we can derive a phenomenological model fit function with three free parameters (Equation 4). This function provides a good fit for the data obtained from diamond anvil cell (d- DAC), laser, and gas gun experiments, as shown in Figure 6.  
+
+<|ref|>text<|/ref|><|det|>[[115, 833, 751, 850]]<|/det|>
+We have removed the assertion: ' since water and ice are incompressible'.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 100, 875, 189]]<|/det|>
+8) It is important to emphasize that the model employed to fit the data is a classical nucleation theory description of the experimental results. It fits the data very well and provides a very good physical picture and interpretation of the experimental results but this depends on parameters used for fitting the data so small changes in the parameters employed could have a significant effect on the model used. This could be mentioned by the authors.  
+
+<|ref|>text<|/ref|><|det|>[[115, 198, 875, 270]]<|/det|>
+We have explicitly stated in the text that the phenomenological model used serves as a fitting function. To further clarify this, we have slightly modified the caption of Figure 6 to read: 'Fitted using the CNT- based phenomenological model given in Equation 4, with the three free parameters equal to ...  
+
+<|ref|>text<|/ref|><|det|>[[115, 303, 878, 410]]<|/det|>
+This report follows up on the previous domain of investigation in d- DAC experiments focusing on the solidification of metastable water. This is the key goal reached in this study. The authors clearly state that support the hypothesis that metastable water, within the pressure range of 1.5 GPa to 2.1 GPa, nucleates into ice VII first, despite this range being the stability field of ice VI. results offer a clarification to the apparent contradiction of the previous findings.  
+
+<|ref|>text<|/ref|><|det|>[[115, 411, 875, 500]]<|/det|>
+In summary, this manuscript clearly addresses and presents a solution to the phase transition discrepancies in the literature. A key improvement in the measurement technique was time- resolved X- ray diffraction. This technique is very well described in the manuscript All figures were clearly presented and definitely provided added clarity to the report. Acceptance is therefore suggested after the minor comments listed above are addressed.  
+
+<|ref|>sub_title<|/ref|><|det|>[[116, 588, 224, 604]]<|/det|>
+## Reviewer #3  
+
+<|ref|>text<|/ref|><|det|>[[115, 622, 870, 890]]<|/det|>
+The authors investigated the interesting behaviour of water ice under dynamic compression with various compression rates using d- DAC by time- resolved x- ray diffraction method and found that ice VII crystallise under the conditions where ice VI should be stable. They also interpret the excess pressure and growth time during the crystallization of ice VII using a phenomenological model based on the classical nucleation theory, succeeding in a comprehensive explanation for the growth rate of ice VII grown under various compression rates. The investigated technique, x- ray diffraction with ruby fluorescence observations at micro- to milli- second order, is somehow remarkable, and the overall achievements of this study would be worth recognising for the ice community. Thus, I do not doubt that this manuscript should be eventually published in at least some specific journal, but I am not sure whether this manuscript will be published by Nature Communication. Two key points in assessing the value of this study would be 1) how new the observed results are and 2) how much their interpretation will influence other related studies. First, concerning the observed results, it should be stated that no truly new phenomena have been reported in this study. It has long been known among ice researchers that ice VII nucleates in the stable region of ice VI even under static compression (e.g. K. Yamamoto, Jpn J Appl Phys,
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 81, 874, 224]]<|/det|>
+19, 1841, 1980, doi: 10.1143/JJAP.19.1841). A recent study using d- DAC has also reported ice VII nucleation in the stability region of ice VI [18]. Therefore, what is new in this manuscript are the results of quantitative experiments on how the growth rate of ice VII changes when the compression rate is varied. The experimental results in this study are novel in the sense that they are quantitative, but the observed phenomena themselves are not particularly novel. The manuscript also does not elaborate on how the quantitative growth rates obtained might affect other related studies. It is therefore difficult to assess, at least at the moment, how important these results are.  
+
+<|ref|>text<|/ref|><|det|>[[115, 233, 866, 305]]<|/det|>
+It is true that the nucleation of ice VII in the stability field of ice VI and the increase in the pressure of overcompressed metastable water with the compression rate were previously known phenomena. But being able to provide an explanation of phenomena through quantitative measurement is the ultimate goal of the physical science research.  
+
+<|ref|>text<|/ref|><|det|>[[115, 313, 864, 525]]<|/det|>
+As acknowledged by reviewer #3, the possibility to control the compression rate over 4 orders of magnitude and to perform time resolved XRD and imaging with the appropriate time- frame from microsecond to millisecond is a significant accomplishment. The quantitative measurements obtained through this approach have enabled us to clarify and unify all previous measurements of the freezing of overcompressed water in a coherent interpretation. Metastable water overcompression under dynamical compression has been the subject of many studies over the past 20 years, many published in high impact journals, as a textbook case of phase transition occurring under far- from- equilibrium conditions. In the present case of water solidification at extreme overcompression, we were able to test a universal scaling law, recently proposed. The Classical Nucleation Theory was shown to unify all previous measurements which lacked the microscopic XRD information.  
+
+<|ref|>text<|/ref|><|det|>[[115, 535, 874, 871]]<|/det|>
+Regarding the solidification of ice VII in the stability domain of ice VI, we thank reviewer#3 for providing this Jp. J. Appl. Phys. reference. However, in it, the reported observation of ice VII in the stability field of ice VI is not exactly what is reported in our study. The nucleation of ice VII was obtained at the metastable extension of the melting line of ice VII in the stability field of ice VI by heating the sample so that the overcompression of water that can be obtained under quasi- static compression could reach the metastable melting line of ice VII. Nucleation of ice VII is stated then to occasionally form and a facetted single crystal of ice VII can bestabilized in equilibrium with the liquid. However, ice VII is kept metastable only if facetted in equilibrium with the liquid. That is exactly what we observed and used in the experiment reported in our supplementary material to precisely extend the ice VII melting line in the stability domain of ice VI. We now have cited the reference in our supplementary information. However, out of this liquid- solid phase equilibrium, by slightly increasing pressure ice VII transforms into ice VI. We are reported a different story here. Under dynamical compression, metastable water always freeze into ice VII in the metastability domain of ice VII and then ice VII transforms to ice VI over a ms time scale and a mixture ice VI - ice VII then exists, which should eventually evolve to an amorphous - ice VI mixture. This resolves all previous contradictions of the literature between reports of amorphous HAD, HAD- ice VI mixture or ice VII.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 100, 877, 152]]<|/det|>
+Even if the importance of this manuscript is recognised by other reviewers and considered for publication in Nat. Commun., the following concerns remain and should be fully considered before potential publication.  
+
+<|ref|>text<|/ref|><|det|>[[115, 170, 866, 294]]<|/det|>
+- The authors assume that the sample temperature should not be increased by the rapid compression, but it should be more carefully evaluated and explained in the manuscript. I also believe that temperature might be quickly stable at room temperature owing to the high thermal conductivity of the diamond. However, if the sample size is enough large, this might not be the case since the thermal conductivity of ice VII would not be such high, so the temperature could temporally increase in a very short time. This issue should be quantitatively assessed, for example, by FEM simulation.  
+
+<|ref|>text<|/ref|><|det|>[[114, 304, 877, 535]]<|/det|>
+The temperature increase in the sample could be an issue. A recent d- DAC measurement leveraged this effect to measure the Gruneisen parameter but for doing that they had to put a layer of zirconia powder on the anvil tip to prevent heat loss from the diamond (ref. 25). We have taken steps to quantify the heating associated to rapid compression by using two pressure luminescence gauges, specifically ruby and SrB40.7:Sm2+, positioned inside the sample chamber, as detailed in ref. 29. The luminescence line of ruby depends on pressure and temperature, whereas that of Samarium depends only on pressure. This method has been previously used to accurately measure melting lines (Datchi et al, PRB 61, 6535). We thus could show that the heating of the sample is negligible. That is illustrated in the figure below showing that, no discernible heating was detected within the measurement's sensitivity, approximately \(\pm 3 \mathrm{~K}\) . Therefore, we can confidently assert that there is no heating occurring. We have now included this quantification of the absence of heating in the supplementary material.
+
+<--- Page Split --->
+<|ref|>image<|/ref|><|det|>[[240, 82, 753, 448]]<|/det|>
+<|ref|>image_caption<|/ref|><|det|>[[120, 453, 875, 480]]<|/det|>
+<center>Dynamic compression of liquid water at a compression rate (10 GPa/ms). The sample pressure over time is measured using simultaneously the \(S\vert \mathrm{text}(SrB)_2[4]\vert \mathrm{text}(O)_2[7]\vert \mathrm{text}(Sm)^\wedge (2 + )S\) and the ruby fluorescence </center>  
+
+<|ref|>text<|/ref|><|det|>[[115, 506, 875, 686]]<|/det|>
+- In the adopted phenomenological model for the nucleation rate, authors ignore the effect of crystal growth, but only take the nucleation rate into account. However, I recognized that the grown ice VII may be somehow coarse crystalline size because the observed diffraction patterns are not like smooth lines but spotty. In particular, I found 111 spots of ice VII, which should be very very weak (generally invisible) if the specimen is fine powder, showing the specimen should be somehow coarse aggregates of many single crystals. The actual rate of ice VII growth may be expressed as the combination of nucleation and growth rates. In fact, the results obtained in this study could be interpreted by different phenomenological models, such as the simple JMAK model. The point would be 'which model is the best' to describe the observed phenomena. Even if one model fits well, this does not necessarily mean that this is the only correct model.  
+
+<|ref|>text<|/ref|><|det|>[[115, 686, 880, 753]]<|/det|>
+Plus, I am also concerned about the lack of temperature terms in the adopted model. I suppose the temperature term should appear since the classical nucleation theory generally has it. Authors should discuss this point (why temperature term could be ignored), if they stick to this model only.  
+
+<|ref|>text<|/ref|><|det|>[[115, 753, 872, 840]]<|/det|>
+First, it is important to note that there is no temperature term involved, as explained above. In the discussion, we begin with the JMAK model and make the assumption that the nucleation rate is predominant over the growth rate. This assumption has been previously utilized in the analysis of overcompression studies of water (ref. 15 and Ref. 16).  
+
+<|ref|>text<|/ref|><|det|>[[115, 850, 877, 904]]<|/det|>
+We have explicitly stated in the text that, for compression rates exceeding a few GPa/ms, this assumption appears reasonable. This conclusion is drawn from the imaging observation of instantaneous nucleation throughout the entire sample chamber, as well
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 82, 850, 152]]<|/det|>
+as the absence of a plateau in the pressure - time evolution during compression. This pertains to the data points presented in figure 6. Conversely, for data points obtained below 1 GPa/ms, a plateau is observed and the growth of the crystallites is imaged, as shown in figure 2. So
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[225, 83, 771, 100]]<|/det|>
+, we believe we correctly explain the framework of our analysis.  
+
+<|ref|>text<|/ref|><|det|>[[115, 126, 880, 181]]<|/det|>
+- Previous studies using d-DAC reported that ice VI and HDA coexist [19,20] as mentioned in the manuscript, but HDA was not observed in this study. The authors should describe the difference from the previous study.  
+
+<|ref|>text<|/ref|><|det|>[[115, 190, 876, 243]]<|/det|>
+At the end of the section on crystallization of ice VII in the stability domain of ice VI, we explicitly mention that ice VII transforming into ice VI could evolve to a mixture of ice VI and HDA. We have no experimental evidence for a more ascertain statement.  
+
+<|ref|>text<|/ref|><|det|>[[115, 262, 223, 278]]<|/det|>
+Minor points  
+
+<|ref|>text<|/ref|><|det|>[[115, 279, 876, 384]]<|/det|>
+1. The precise wavelength should be described in Method section. I found the \(x\) -ray energy is \(19 \mathrm{keV}\) , but the number of significant figures is only two in this case.  
+2. The reflection indices should not be in parenthesis. Index in parenthesis, (hkl) means Miller index that is used for corresponding Miller plane in real space. But reflection index, hkl, in reciprocal space should not have any parenthesis (see International Tables for crystallography, for this criterion).  
+
+<|ref|>text<|/ref|><|det|>[[115, 394, 483, 411]]<|/det|>
+The manuscript was modified accordingly.  
+
+<|ref|>text<|/ref|><|det|>[[115, 439, 864, 474]]<|/det|>
+3. Effective digit and/or the error should be considered throughout the whole manuscript, in particular, in Table I.  
+
+<|ref|>text<|/ref|><|det|>[[115, 484, 740, 502]]<|/det|>
+This comment was taken into account throughout the whole manuscript.
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[118, 85, 316, 101]]<|/det|>
+## REVIEWER COMMENTS  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 137, 437, 154]]<|/det|>
+## Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[118, 163, 834, 206]]<|/det|>
+The authors have adequately responded to my comments on their first submission. I'm happy to recommend publication.  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 266, 437, 284]]<|/det|>
+## Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[118, 292, 792, 336]]<|/det|>
+This manuscript has clarified the occurrence of crystallization into ice VII occurs in between19  
+
+<|ref|>text<|/ref|><|det|>[[118, 344, 866, 414]]<|/det|>
+1.6 GPa and 2.0 GPa, that is in the stability field of ice VI. This was achieved by employing time-resolved X-ray diffraction and a Diamond-Annil-Cell, with pressure rise times from 0.1 ms to 100 ms.  
+
+<|ref|>text<|/ref|><|det|>[[117, 422, 872, 492]]<|/det|>
+The authors have added appropriate references as suggested by the reviewers. The authors have also added appropriate clarifying sentences in response to reviewer's remarks and questions.  
+
+<|ref|>text<|/ref|><|det|>[[117, 500, 878, 623]]<|/det|>
+An important exception to this is an incorrect comment made by the authors regarding an important reference to an article by cited by a reviewer. The authors state in their rebuttal comments "We now have cited the Yamamoto reference in our supplementary information" This is not seen in the revised manuscript and supplementary information copy submitted by the authors.  
+
+<|ref|>text<|/ref|><|det|>[[117, 631, 850, 728]]<|/det|>
+In summary, this revised manuscript describes in more detail than previously published studies the results of a re- examination of the phase transition of water to ice VII. It is not clear however that providing additional details of a known phase transition qualifies for acceptance in Nature.  
+
+<|ref|>sub_title<|/ref|><|det|>[[118, 789, 437, 805]]<|/det|>
+## Reviewer #3 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[118, 815, 415, 831]]<|/det|>
+Please find the attached comments.  
+
+<|ref|>text<|/ref|><|det|>[[118, 850, 719, 868]]<|/det|>
+[editorial note: please see the next page(s) for reviewer #3's comments.]
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[87, 87, 900, 128]]<|/det|>
+Comments on the manuscript entitled 'Metastable water at several compression rates and its freezing kinetics into ice VII" by Pepin et al.,  
+
+<|ref|>text<|/ref|><|det|>[[87, 175, 880, 263]]<|/det|>
+I noticed other two reviewers appreciated the importance of this study, which is in itself pleasing as a researcher in ice study. Although I understand most of authors rebuttal comments, authors replies to my comments raised the following further concerns. I point out my concerns as replies (bold) following to the rebuttal comments from the authors (italic).  
+
+<|ref|>text<|/ref|><|det|>[[88, 311, 104, 325]]<|/det|>
+1.  
+
+<|ref|>text<|/ref|><|det|>[[85, 342, 911, 568]]<|/det|>
+Regarding the solidification of ice VII in the stability domain of ice VI, we thank reviewer#3 for providing this Jp. J. Appl. Phys. reference. However, in it, the reported observation of ice VII in the stability field of ice VI is not exactly what is reported in our study. The nucleation of ice VII was obtained at the metastable extension of the melting line of ice VII in the stability field of ice VI by heating the sample so that the overcompression of water that can be obtained under quasi- static compression could reach the metastable melting line of ice VII. Nucleation of ice VII is stated then to occasionally form and a facetted single crystal of ice VII can bestabilized in equilibrium with the liquid. However, ice VII is kept metastable only if facetted in equilibrium with the liquid. That is exactly what we observed and used in the experiment reported in our supplementary material to precisely extend the ice VII melting line in the stability domain of ice VI. We now have cited the reference in our supplementary information.  
+
+<|ref|>text<|/ref|><|det|>[[86, 581, 902, 715]]<|/det|>
+However, out of this liquid- solid phase equilibrium, by slightly increasing pressure ice VII transforms into ice VI. We are reported a different story here. Under dynamical compression, metastable water always freeze into ice VII in the metastability domain of ice VII and then ice VII transforms to ice VI over a ms time scale and a mixture ice VI - ice VII then exists, which should eventually evolve to an amorphous - ice VI mixture. This resolves all previous contradictions of the literature between reports of amorphous HAD (could be HDA?), HAD- ice VI mixture or ice VII.  
+
+<|ref|>text<|/ref|><|det|>[[88, 763, 340, 779]]<|/det|>
+and the reply to the last comment.  
+
+<|ref|>text<|/ref|><|det|>[[85, 794, 905, 890]]<|/det|>
+- Previous studies using d-DAC reported that ice VI and HDA coexist [19,20] as mentioned in the manuscript, but HDA was not observed in this study. The authors should describe the difference from the previous study. At the end of the section on crystallization of ice VII in the stability domain of ice VI, we explicitly mention that ice VII transforming into ice VI could evolve to a mixture of ice VI and HDA. We have no experimental
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[87, 186, 896, 342]]<|/det|>
+I agree the dynamic compression may be somehow different to the static compression, but one point I would indicate is that the metastable ice VII made by the static compression in ice VI stability region also eventually convert to ice VI, since it is metastable. The difference to the static compression would be the degree of over- compression (AG between liquid and ice VII). Facet morphology found in ice VII by the static compression could also be found in that made by the dynamic compression in sub- micrometer scale, which could be just the difference of size. In that sense, the phenomenological difference between static and dynamic compression may be still unclear.  
+
+<|ref|>text<|/ref|><|det|>[[86, 355, 907, 536]]<|/det|>
+In this study, the HDA- ice VI mixture was not observed, and may not be explained the clear reason why HDA was observed in the previous study as stated in the reply to the last comment. It is briefly mentioned in the line 190- 197, as 'This difference could be explained by the fact that at 1.8 GPa, ice VII is approaching its thermodynamic stability, allowing for the stabilization of a mixture of ice VI and ice VII, whereas at 1.7 GPa, ice VII is gradually transforming, leading to the stabilization of a mixture of ice VI and amorphous ice.' But I could not satisfy this description resolves all previous contractions. The present results would show even another contradiction in the sense that they could not observe amorphous state.  
+
+<|ref|>text<|/ref|><|det|>[[87, 583, 104, 598]]<|/det|>
+2.  
+
+<|ref|>text<|/ref|><|det|>[[87, 614, 907, 702]]<|/det|>
+We thus could show that the heating of the sample is negligible. That is illustrated in the figure below showing that, no discernible heating was detected within the measurement's sensitivity, approximately \(\pm 3 \text{K}\) . Therefore, we can confidently assert that there is no heating occurring. We have now included this quantification of the absence of heating in the supplementary material.
+
+<--- Page Split --->
+<|ref|>image<|/ref|><|det|>[[100, 90, 410, 280]]<|/det|>
+<|ref|>image_caption<|/ref|><|det|>[[92, 281, 464, 295]]<|/det|>
+<center>Dynamic compression of liquid water at a compression rate (10 GPa/ms). The sample pressure over time is measured using simultaneously the S(Text(Si6)_[4]/text(O)_[7]/text(Sm)^(2+)/S and the ruby fluorescence </center>  
+
+<|ref|>text<|/ref|><|det|>[[87, 342, 901, 430]]<|/det|>
+From the above figure, I could see significant increases in temperatures, exceeding 315 K (42 °C), it does not look like \(\pm 3\) K. Plus, the plot in Fig. 9 shows the temperature \(\sim 21\) °C (294 K), it contradicts with the above figure. If the temperature changes over 10 K, the story might change since the experimental conditions are close to the melting line.  
+
+<|ref|>image<|/ref|><|det|>[[108, 455, 603, 792]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[87, 852, 755, 869]]<|/det|>
+Finally one minor question for the above figure, what is the blue atom in ice VI structure?
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[115, 84, 528, 102]]<|/det|>
+## Detailed answer to reviewers' comments.  
+
+<|ref|>sub_title<|/ref|><|det|>[[115, 130, 248, 148]]<|/det|>
+## Reviewer #1:  
+
+<|ref|>text<|/ref|><|det|>[[115, 167, 856, 199]]<|/det|>
+The authors have adequately responded to my comments on their first submission. I'm happy to recommend publication.  
+
+<|ref|>text<|/ref|><|det|>[[115, 227, 505, 243]]<|/det|>
+We thank reviewer #1 for his recommendation.  
+
+<|ref|>sub_title<|/ref|><|det|>[[115, 288, 248, 306]]<|/det|>
+## Reviewer #2:  
+
+<|ref|>text<|/ref|><|det|>[[115, 324, 875, 390]]<|/det|>
+This manuscript has clarified the occurrence of crystallization into ice VII occurs in between 1.6 GPa and 2.0 GPa, that is in the stability field of ice VI. This was achieved by employing time- resolved X- ray diffraction and a Diamond- Anvil- Cell, with pressure rise times from 0.1 ms to 100 ms.  
+
+<|ref|>text<|/ref|><|det|>[[115, 392, 876, 492]]<|/det|>
+The authors have added appropriate references as suggested by the reviewers. The authors have also added appropriate clarifying sentences in response to reviewer's remarks and questions. An important exception to this is an incorrect comment made by the authors regarding an important reference to an article by cited by a reviewer. The authors state in their rebuttal comments "We now have cited the Yamamoto reference in our supplementary information" This is not seen in the revised manuscript and supplementary information copy submitted by the authors.  
+
+<|ref|>text<|/ref|><|det|>[[115, 493, 866, 544]]<|/det|>
+In summary, this revised manuscript describes in more detail than previously published studies the results of a re- examination of the phase transition of water to ice VII. It is not clear however that providing additional details of a known phase transition qualifies for acceptance in Nature.  
+
+<|ref|>text<|/ref|><|det|>[[115, 562, 848, 612]]<|/det|>
+We thank reviewer #2 for his feedback on our revised manuscript. We acknowledge the oversight regarding the the Yamamoto reference and have now included it in the revised supplementary information.  
+
+<|ref|>text<|/ref|><|det|>[[115, 629, 876, 694]]<|/det|>
+However, we are surprised by the change in Reviewer #2's appreciation of our manuscript. Initially, the reviewer suggested acceptance after minor comments were addressed, but he/she is now questioning the suitability of our work for publication in Nature. This change in stance is unexpected and lacks clarification.  
+
+<|ref|>sub_title<|/ref|><|det|>[[115, 763, 248, 781]]<|/det|>
+## Reviewer #3:  
+
+<|ref|>text<|/ref|><|det|>[[115, 820, 866, 870]]<|/det|>
+I noticed other two reviewers appreciated the importance of this study, which is in itself pleasing as a researcher in ice study. Although I understand most of authors rebuttal comments, authors replies to my comments raised the following further concerns  
+
+<|ref|>text<|/ref|><|det|>[[115, 881, 845, 914]]<|/det|>
+I agree the dynamic compression may be somehow different to the static compression, but one point I would indicate is that the metastable ice VII made by the static compression in ice VI
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 81, 877, 301]]<|/det|>
+stability region also eventually convert to ice VI, since it is metastable. The difference to the static compression would be the degree of over- compression (AG between liquid and ice VII). Facet morphology found in ice VII by the static compression could also be found in that made by the dynamic compression in sub- micrometer scale, which could be just the difference of size. In that sense, the phenomenological difference between static and dynamic compression may be still unclear. In this study, the HDA- ice VI mixture was not observed, and may not be explained the clear reason why HDA was observed in the previous study as stated in the reply to the last comment. It is briefly mentioned in the line 190- 197, as 'This difference could be explained by the fact that at 1.8 GPa, ice VII is approaching its thermodynamic stability, allowing for the stabilization of a mixture of ice VI and ice VII, whereas at 1.7 GPa, ice VII is gradually transforming, leading to the stabilization of a mixture of ice VI and amorphous ice.' But I could not satisfy this description resolves all previous contractions. The present results would show even another contradiction in the sense that they could not observe amorphous state.  
+
+<|ref|>text<|/ref|><|det|>[[115, 311, 878, 532]]<|/det|>
+We respectfully disagree with reviewer #3. There is a phenomenological difference between static and dynamic compressions concerning the observation of stable ice VII in the stability field of ice VI. In the static experiment, ice VII was brought into the stability field of ice VI by following the metastable melting line of ice VII down to 300K. If the melting equilibrium is lost and complete solidification occurs, ice VII transforms to ice VI. In dynamical compression, the over- compression pressure of metastable water is increasing with the compression rate. For compression rates in between 0.15 GPa/ms sans 0.9 GPa/ms, the freezing of metastable water falls in the stability domain of ice VI and in the metastability domain of ice VII, as determined by the extension of the melting line of ice VII. Our study shows that metastable water under such conditions always freezes in ice VII and then transforms into ice VI over a few milliseconds. Using time resolved XRD we could thus show a preferred nucleation mechanism into ice VII which was not shown by static measurements.  
+
+<|ref|>text<|/ref|><|det|>[[115, 541, 878, 763]]<|/det|>
+Regarding the formation of amorphous ice, HDA, as reported by previous studies based on Raman measurements, we are unable to observe it because it is a too weak signal by X- ray diffraction. It should also be noted that a mixture of fine powder of ice VI and ice VII would have a Raman signal that looks like the one of HDA ice. The important points of our study are the preferred nucleation of water into ice VII and then the time- scale of the transformation into ice VI. Whether there is complete transformation into ice VI, a mixture of ice VI and ice VII, a mixture of ice VI and HDA could depend of the final pressure of the rapid compression. We have modified the manuscript to make this point clearer. The sentence now reads: "This difference could be explained by the fact that at 1.8 GPa, ice VII is approaching its thermodynamic stability, allowing for the stabilization of a mixture of ice VI and ice VII (and a possible amorphous state, which could not be observed in our case due to its very low diffraction power), whereas at 1.7 GPa, ice VII has been reported to gradually transform in a mixture of ice VI and amorphous ice."  
+
+<|ref|>text<|/ref|><|det|>[[116, 799, 864, 866]]<|/det|>
+From the above figure, I could see significant increases in temperatures, exceeding 315 K (42 °C), it does not look like \(\pm 3\) K. Plus, the plot in Fig. 9 shows the temperature \(\sim 21\) °C (294 K), it contradicts with the above figure. If the temperature changes over 10 K, the story might change since the experimental conditions are close to the melting line.
+
+<--- Page Split --->
+<|ref|>image<|/ref|><|det|>[[228, 82, 765, 460]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[115, 475, 822, 527]]<|/det|>
+We apologize for any confusion caused by the previous error bars in our temperature measurements. The \(\pm 3 \mathrm{K}\) error bar was taken from the reference article based on the luminescence gauges, but we have since re- estimated the error bars to be \(\pm 10 \mathrm{K}\) .  
+
+<|ref|>text<|/ref|><|det|>[[115, 536, 880, 688]]<|/det|>
+We would like to point out that the temperature relation in our study was obtained in non- hydrostatic pressure conditions, whereas in ref. 29, the error bar estimation was obtained in a hydrostatic pressure medium. This impacts the fluorescence of the ruby by broadening the peaks and making them asymmetrical, which slightly skews the pressure measurements. This is evidenced in the bottom graph of the figure provided. If a temperature increase were to take place during the fast compression, the temperature should be an increasing function of time throughout the entire compression sequence. However, this is not the case here, as we observed the temperature oscillating around the initial temperature of \(294 \mathrm{K}\) , within a \(\pm 10 \mathrm{K}\) error bar.  
+
+<|ref|>text<|/ref|><|det|>[[115, 699, 863, 766]]<|/det|>
+Furthermore, if we consider a temperature increase of \(10 \mathrm{K}\) for our point under 14.82 GPa/ms, the figure below shows that it corresponds to a pressure difference in the overcompression of metastable water of \(\sim 0.06 \mathrm{GPa}\) , which is within the error bars that we considered in our models of figures 5 and 6.  
+
+<|ref|>text<|/ref|><|det|>[[115, 777, 866, 810]]<|/det|>
+In conclusion, the temperature increase during compression is very small and if any, of the order of \(10 \mathrm{K}\) , does not affect the overall results and conclusions of our study.  
+
+<|ref|>text<|/ref|><|det|>[[115, 820, 875, 871]]<|/det|>
+We have slightly modified the manuscript to reflect this and the figure as shown above. The sentence now reads: "That is illustrated by the figure below showing that, no discernible heating was detected within the measurement's sensitivity, approximately \(\pm 10 \mathrm{K}\) ".
+
+<--- Page Split --->
+<|ref|>image<|/ref|><|det|>[[212, 85, 780, 475]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[115, 536, 816, 553]]<|/det|>
+Finally one minor question for the above figure, what is the blue atom in ice VI structure?  
+
+<|ref|>text<|/ref|><|det|>[[115, 562, 848, 597]]<|/det|>
+This is just a visual representation of the oxygen atoms occupying two different positions (Wycoff site 2b in blue and 8g in red).
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[119, 85, 330, 101]]<|/det|>
+## REVIEWERS' COMMENTS  
+
+<|ref|>text<|/ref|><|det|>[[118, 137, 439, 180]]<|/det|>
+Reviewer #3 (Remarks to the Author): Please find the attached comments.  
+
+<|ref|>text<|/ref|><|det|>[[118, 224, 720, 241]]<|/det|>
+[editorial note: please see the next page(s) for reviewer #3's comments.]
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[88, 84, 909, 133]]<|/det|>
+1. I acknowledge the author's comments on my previous concerns regarding the phenomenological difference between static and dynamic compression. First, I would indicate one fact that ice VII could also nucleate into the stability field of ice VI as written by Yamamoto et al., described as  
+
+<|ref|>text<|/ref|><|det|>[[87, 142, 904, 176]]<|/det|>
+"...it was discovered in the present experiment that ice VII occasionally formed directly from water at point B in Fig. 1 without the formation of ice VI at point A." (see below)  
+
+<|ref|>text<|/ref|><|det|>[[295, 330, 439, 347]]<|/det|>
+[figure redacted]  
+
+<|ref|>text<|/ref|><|det|>[[88, 686, 543, 703]]<|/det|>
+Therefore, the author's indication in their rebuttal comments,  
+
+<|ref|>text<|/ref|><|det|>[[88, 711, 899, 760]]<|/det|>
+"In the static experiment, ice VII was brought into the stability field of ice VI by following the metastable melting line of ice VII down to 300K. If the melting equilibrium is lost and complete solidification occurs, ice VII transforms to ice VI."  
+
+<|ref|>text<|/ref|><|det|>[[88, 770, 243, 785]]<|/det|>
+is misunderstanding.  
+
+<|ref|>text<|/ref|><|det|>[[88, 794, 861, 844]]<|/det|>
+But, if in the dynamic compression ice VII "always" preferably nucleates rather than ice VI, this is the phenomenological difference since in the static compression ice VII could only "occasionally" nucleate. Authors could emphasize this point with the reference to Yamamoto et al.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[87, 83, 907, 215]]<|/det|>
+2. As for the description of (non)appearance of HDA, the added sentence "(and a possible amorphous state, which could not be observed in our case due to its very low diffraction power)" is a bit inaccurate since the diffraction power ("scattering power" may be more appropriate term in the case of amorphous, though "diffraction power" is also correct) itself is comparable to the crystalline material if all integrated through whole reciprocal space. I suggest this sentence would be "...in our case since the scattering from amorphous may be hidden in the Bragg peaks even it exists" or so. In any case, the last sentence of this paragraph, "These results offer a clarification to the apparent contradiction of the previous findings." seems a bit overstatement to me.  
+
+<|ref|>text<|/ref|><|det|>[[87, 224, 907, 338]]<|/det|>
+3. I consider, repeatedly saying, that the difference to static compression would be the degree of overcompression. The over-compression as large as a few GPa could not be achieved by the static compression. The strong point of this study would be the quantitative analysis of the nucleation time as a function of the over-compression. This is the reason why I repeatedly indicate the quantitative considerations for the temperature effect since I suppose the \(10\mathrm{K}\) difference is huge for water molecules at a few GPa regions. I suggest some comments on the temperature effect, as written in the rebuttal comments, may be included in the main text as well.  
+
+<|ref|>text<|/ref|><|det|>[[88, 346, 622, 363]]<|/det|>
+4. Regarding the qualitative analysis, I pointed out in the first comment,  
+
+<|ref|>text<|/ref|><|det|>[[87, 371, 902, 404]]<|/det|>
+"Effective digit and/or the error should be considered throughout the whole manuscript, in particular, in Table I."  
+
+<|ref|>text<|/ref|><|det|>[[88, 414, 235, 430]]<|/det|>
+and authors replied,  
+
+<|ref|>text<|/ref|><|det|>[[88, 439, 640, 456]]<|/det|>
+"This comment was taken into account throughout the whole manuscript."  
+
+<|ref|>text<|/ref|><|det|>[[87, 464, 910, 580]]<|/det|>
+I did not carefully check in the second round (since I believed this reply), but I realized that Table I was not modified yet. For example, the rate of compression in the first raw is written as '870787.5 (GPa/ms)". I could not believe the effective digit could reach 7 digits, such very precise pressure control would not be possible. Similarly, in the caption of Figure 6, the parameter b is written like "b=245293534.93", which is also unrealistic effective digit considering the fitting result shown in Figure 6. Such representations without care of effective digits led readers to the question of the whole results of this study. Authors should carefully check again this point.
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[116, 84, 528, 103]]<|/det|>
+## Detailed answer to reviewers' comments.  
+
+<|ref|>text<|/ref|><|det|>[[115, 136, 870, 201]]<|/det|>
+1- I acknowledge the author's comments on my previous concerns regarding the phenomenological difference between static and dynamic compression. First, I would like to point out that ice VII could also nucleate within the stability field of ice VI, as written by Yamamoto et al., described as:  
+
+<|ref|>text<|/ref|><|det|>[[115, 202, 876, 366]]<|/det|>
+"...it was discovered in the present experiment that ice VII occasionally formed directly from water at point B in Fig. I without the formation of ice VI at point A." (see below) Therefore, the author's statement in their rebuttal comments, "In the static experiment, ice VII was brought into the stability field of ice VI by following the metastable melting line of ice VII down to 300K. If the melting equilibrium is lost and complete solidification occurs, ice VII transforms to ice VI." is a misunderstanding. However, if under dynamic compression ice VII "always" nucleates preferentially rather than ice VI, this would be the phenomenological difference, since under static compression ice VII could only "occasionally" nucleate. The authors could emphasize this point with reference to Yamamoto et al.  
+
+<|ref|>text<|/ref|><|det|>[[116, 382, 879, 434]]<|/det|>
+Following the reviewer's advice, we have added a sentence that reads: "It was also shown that ice VII could occasionally nucleate in the stability domain of ice VI above the metastable melting line [41]".  
+
+<|ref|>text<|/ref|><|det|>[[115, 448, 867, 614]]<|/det|>
+2- Regarding the description of the (non)appearance of HDA, the added sentence "(and a possible amorphous state, which could not be observed in our case due to its very low diffraction power)" is somewhat inaccurate, as the diffraction power ("scattering power" might be a more appropriate term in the case of amorphous materials, although "diffraction power" is also correct) itself is comparable to that of crystalline material if integrated over the whole reciprocal space. I suggest rephrasing this sentence as "...in our case since the scattering from amorphous material may be hidden in the Bragg peaks even if it exists," or something similar. In any case, the last sentence of this paragraph, "These results offer a clarification to the apparent contradiction of the previous findings," seems a bit of an overstatement to me.  
+
+<|ref|>text<|/ref|><|det|>[[116, 629, 480, 646]]<|/det|>
+We have modified the main text accordingly.  
+
+<|ref|>text<|/ref|><|det|>[[115, 661, 876, 778]]<|/det|>
+3- I believe, and have repeatedly stated, that the difference with static compression lies in the degree of over-compression. Over-compression as large as a few GPa could not be achieved by static compression. The strength of this study lies in the quantitative analysis of nucleation time as a function of over-compression. This is why I repeatedly emphasize the importance of quantitative considerations regarding the temperature effect, as I believe the 10 K difference is significant for water molecules under a few GPa. I suggest that some comments on the temperature effect, as written in the rebuttal comments, be included in the main text as well.  
+
+<|ref|>text<|/ref|><|det|>[[115, 792, 878, 910]]<|/det|>
+We have added a statement in the supplementary information. It reads: "We would like to point out that the temperature relation in our study was obtained in non- hydrostatic pressure conditions, whereas in Ref.\cite{Datchi97}, the error bar estimation was obtained in a hydrostatic pressure medium. This impacts the fluorescence of the ruby by broadening the peaks and making them asymmetrical, which slightly skews the pressure measurements. This is evidenced in the bottom graph of Supplementary Figure \ref{fig:figureS4}. If a temperature increase were to take place during the fast compression, the temperature should be an
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 82, 876, 132]]<|/det|>
+increasing function of time throughout the entire compression sequence. However, this is not the case here, as we observed the temperature oscillating around the initial temperature of 294 K, within a \(\pm 10 \mathrm{K}\) error bar.  
+
+<|ref|>text<|/ref|><|det|>[[115, 148, 872, 214]]<|/det|>
+Furthermore, if we consider a temperature increase of \(10 \mathrm{K}\) for the point at \(14.82 \mathrm{GPa / ms}\) , it would correspond to a pressure difference in the overcompression of metastable water of \(- 0.06 \mathrm{GPa}\) , which is within the error bars that we considered in our models of figures 5 and 6. ”  
+
+<|ref|>text<|/ref|><|det|>[[115, 231, 870, 428]]<|/det|>
+4- Regarding the qualitative analysis, I pointed out in my first comment, "Effective digits and/or errors should be considered throughout the entire manuscript, particularly in Table I," and the authors replied, "This comment was taken into account throughout the entire manuscript." I did not carefully check this during the second round (as I trusted their reply), but I later realized that Table I had not yet been modified. For example, the rate of compression in the first row is written as '870787.5 (GPa/ms).' I cannot believe the effective digits could reach seven digits; such precise pressure control would not be possible. Similarly, in the caption of Figure 6, the parameter \(b\) is written as " \(b = 245293534.93\) ," which is also an unrealistic number of significant digits considering the fitting result shown in Figure 6. Such representations without careful attention to significant digits lead readers to question the overall results of this study. The authors should carefully check this point again.  
+
+<|ref|>text<|/ref|><|det|>[[115, 444, 880, 494]]<|/det|>
+We have modified the main text accordingly, carefully checking for effective digits. However, some number have not been modified because they are derived from data from references cited in our manuscript.
+
+<--- Page Split --->

peer_reviews/1287b0275a1ceb77692926e47e4cb244d3ebcdcf9220bb8cc464a7bb7c810ed0/supplementary_0_Peer Review File/images_list.json

@@ -0,0 +1,10 @@
+[
+  {
+    "type": "image",
+    "img_path": "images/Supplementary_Figure_12.jpg",
+    "caption": "Rebuttal Figure 1 Posterior probability functions of the inverted source parameters from Event 2. Notations are the same as Supplementary Figure 12 in the manuscript.",
+    "footnote": [],
+    "bbox": [],
+    "page_idx": 0
+  }
+]

peer_reviews/1287b0275a1ceb77692926e47e4cb244d3ebcdcf9220bb8cc464a7bb7c810ed0/supplementary_0_Peer Review File/supplementary_0_Peer Review File.mmd

@@ -0,0 +1,947 @@
+
+# nature portfolio  
+
+Peer Review File  
+
+Episodic transport of discrete magma batches beneath Aso volcano  
+
+![](images/Supplementary_Figure_12.jpg)
+  
+
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+
+## REVIEWER COMMENTS  
+
+Reviewer #1 (Remarks to the Author):  
+
+## Overall impression  
+
+Overall impressionThe article written by Niu and Song shows an exciting interpretation of tons of tilt and seismic signals that repeatedly occurred at Aso volcano, Japan, with precise analysis procedures and its results. It sheds light on the new concept of the magma plumbing system at the volcano, adding a magma storage zone located between the magma chamber and the conduit. It identifies that this zone's inflation and deflation are associated with episodic magma discharge from below and are indeed relating to the surficial phenomena. After calculating the 1- D conduit flow, it improves the public image of magma ascent dynamics from the magma chamber to the surface. It also provides a future direction of geophysical research to real- time evaluate upcoming eruptions. This article seems to be well- organized, and the methods and results are detailed and very interesting for the scientific community widely. I would recommend their article for publishing in the journal after minor revisions of some comments below I pointed out.  
+
+## Specific comments  
+
+P3 L13 - The authors use the term "long- period tremor (LPT)" for the objected seismic signal in the whole in this article because they respect the nomenclature that had been established in previous papers. However, this term seems to be a local vernacular or jargon only at Aso volcano. I am also suspect that that signal is not a "tremor?" I suggest using the more widely used term VLP or ultra- long- period (ULP) signals in the volcanological field instead of LPT for readers of this journal from more wide backgrounds.  
+
+P4 L4 - Were the VLP signals the authors checked from the catalog associated individually with any surficial eruption activities at the crater during the erupted periods? A paper reported that the occurrence of the VLP seismic signal preceded by few seconds from the onset of Strombolian eruption (Ishii et al., 2019, EPS). What the meaning of a difference of such elapsed time (2 min for Oct. 2016 phreatomagmatic eruption and 2- 3 s for Strombolian explosions), as well as the meaning of a difference with or without eruptions? I think such a time difference may not relate to the patterns of the tilt source. Are there several types of VLPs with similar waveforms, but specific properties are different? Can the author comment on this topic at any part of this article?  
+
+P4 L20 - Can the authors show a result of the same method using another template (Event 2) in the supplementary? I could not find validity to use Event 1 (not 2) as a reference signal in this article. I am convincing that the conclusion will be the same as this article if Event 2 is used. However, curious about the time evolution of the VLPs for the case of 2.  
+
+P5 L15 - The authors should explain the excitation mechanism of the SPT concerning the occurrence of VLP as well as the series of phenomena starting from a deeper place (top of the magma chamber; the authors argue). I could not imagine why the SPT starts at the top portion of the conduit (VLP source) at first, and 10 s later, the VLP occurs (it is the same time as the SPT peak), whenever the events are inflation or deflation.  
+
+P6 L10 - The authors cited Hata et al. (2018, JGR) in this article; however, it seems inappropriate. At least another paper of Hata et al. (2018, JGR, 10.1029/2018JB015951), or much preferable Matsushima et al. (2020, EPS) showing the revised model of the Hata et al.'s result should be cited.  
+
+P6 L20 - Add any comment on how to make seismicity around the roof of the magma chamber if gas- dominant materials transport upward.  
+
+P7 L1 - Can the authors show other evidence of inflation and deflation during these periods? It would be enough to cite some papers of InSAR or GNSS that can strengthen their argument.  
+
+P7 L14 - I could not understand the concept relating to SO2 gas emission. Could the author explain more carefully? The authors describe the prominence of inflation events equivalent to pressurization and lowering SO2 emission. The relation between inflation and pressurization is readily accepted because both are relative
+
+<--- Page Split --->
+
+changes of the SMSZ condition. However, in my understanding, SO2 emission does not seem to this relative change; but relates to the absolute volume of magma transported to a shallow depth. Is this wrong? I could not see a significant correlation between Fig 4b and 4c.  
+
+P9 L11 - Probably the wrong citation; did not Miyabuchi & Hara (2019, EPS) treat the 2016 phreatomagmatic explosion? I guess Ishii (2018, EPS), Ishii et al. (2018, EPS) and Sato et al. (2018, EPS) are more suitable for the mass of ejected materials.  
+
+P11 L 13 - Unclear to me.  
+
+References - It is not good to cite several papers written in Japanese, which most of the readers cannot probably reach and read. Other accessible papers should be referred to in this article as far as possible. The author also should discuss the proposed plumbing system with the result of Tsutsui & Sudo (2004, JVGR).  
+
+Fig. 2 - I guess the exact time of the onset in longer signals is quite tricky. How about the reading error?  
+
+P37 L 14 - need a reference for assumption (or evaluation)  
+
+Reviewer #2 (Remarks to the Author):  
+
+This paper provides interesting new observations of coupled seismic and ground deformation of repeated magma transport events observed at Aso volcano, Japan that are at least sometimes associated with eruptions. These types of high spatial and temporal resolution observations combining these datasets are rare (to my knowledge) and provide a unique perspective on the timing, location, and volume change of magma movements within a volcano. It would be good to get the perspective of an expert on the Aso system, but as far as I can tell, this paper provides new insight into the plumbing system at this volcano and is possibly applicable (in terms of the conceptual model and technique) to other volcanic systems.  
+
+I recommend publication after moderate revision. There are several steps in the analysis that are not well documented (see detailed comments below). Further, the paper is unclear in some locations or the discussion is incomplete (again documented below).  
+
+Page 1:  
+
+Line 10 and Page 1 lines 6- 9: How do we know that Aso has a crystal rich mush and that it is relevant for this study? Maybe the magma batches are coming from a crystal poor reservoir? Maybe the crystal rich mush is deeper?  
+
+Line 14: "individual" should be "an individual"  
+
+Lines 16- 24: I found the following sentences unclear and confusing - - can the authors be more specific? What do you mean by "composition dependent"? Is the composition of each magma batch different? Does the last sentence mean that you can forecast the eruption style, plume height, etc. based on the tilt/seismic data described above? If so provide some more details as to how.  
+
+"whereas their recurrences, potentially composition dependent, are regulated by the brittle- to- ductile transition rheology under low differential stress and high strain rate due to the surge of magma from below, regulating long- term volcanic output rate. The magma ascent velocity, decompression rates, and cumulative magma output deduced from the episodic deformation events before recent eruptions in Aso volcano are compatible with retrospective observations of the eruption style, tephra fallouts, and plume heights, promising real- time evaluation of upcoming eruptions."  
+
+Further, the results shown in Figures 4- 6 aren't really described in the abstract.
+
+<--- Page Split --->
+
+Page 2:  
+
+Line 5: the ambient stress state also matters  
+
+Line 16: Also the ambient stress state matters - - seismicity will only occur where the rocks are near to failure. Magma can move aseismically if the stress state is not close to failure.  
+
+Page 3:  
+
+Line 8: "signal" should be "signals"  
+
+Line 10: How do we know this is a "shallow hydrothermal reservoir"?  
+
+Line 11: use "on" instead of "against"  
+
+Line 12 (and Page 1 line 12): Mentioning the source is near sea level is confusing, how far is this below the surface? It would be better to tell us the depth of the source beneath the surface (or at least tell us both pieces of information).  
+
+Line 16: Need some introduction to the eruptive cycle - - why was the time period 2011- 2016 chosen? Why not a longer time period? Is the LPT only seen in this time period? Is this the only time period when the patterns described below occur? Or some other reason?  
+
+Line 17: "waveform" should be "waveforms"  
+
+Line 21: How do you know these LPT events are "anomalous"? Where do you define normal or background LPT activity?  
+
+Page 4:  
+
+Line 4: where do the displacement waveforms come from? Integration of the seismograms? If so, how?  
+
+Line 5: are these events associated with eruptions?  
+
+Line 6: What does "east- down" mean? Doe that mean tilt toward the east?  
+
+Line 8- 9: This phrase could be more precise: "between the signal of LPT and the tilt offset". Perhaps: "between the LPT signal and the tilt offset at different stations and in the different components at the same station."  
+
+Line 17: Are all the LPT events associated with eruptions? Are there LPT events that aren't found by the matched filter? If so, what type of events do they represent?  
+
+The sections entitled "LPT and synchronous tilt/displacement offset" and "Discovery of the inflation/deflation event beneath Aso volcano" could be better organized.  
+
+It seems to organized in a chronological manner instead of a logical description of what was discovered - - the first paragraph talks about the 2016 eruption, the next paragraph is about a manual search and the next paragraph is about a matched filter. Instead, why not just discuss the procedure (pointing to the Materials and Methods as needed) and then describe what you found? Maybe organize: this is what we analyzed, describe the 2016 events (including the variation in signals between stations) and then the global stack.  
+
+Page 5,
+
+<--- Page Split --->
+
+line 4: How many events are in the stack? Do the number of events vary in time in a systematic manner? What do the unstacked events look like?  
+
+Line 5: what does volcanic unrest mean here? Are there eruptions in 2011- 2014 or are these events occurring without eruption? If so, that is strange - - why do these similar events occur sometimes with eruption and sometimes without?  
+
+Page 6:  
+
+Line 4: Make clear from the source what your forward model is in terms of source characteristics and elastic structure.  
+
+Line 15: What is the composition of the magma batch based on the eruption? One hypothesis is that this portion of magma is ascending because it has accumulated enough gas to become buoyant, so what is known about the gas compositions in the eruptions?  
+
+Line 16: How do we know that this is a crystal rich magma? Is this just a guess or is there some evidence from petrology about the source region?  
+
+Line 20: An alternative is that the magma is that there is no new injection of magma and it is cooling/crystallizing, accumulating gas at the top of the reservoir and then episodically having sufficient buoyancy to cause brittle failure. (This is a top- down instead of a bottom- up trigger for eruptions, see for example Girona, T., Costa, F., & Schubert, G. (2015). Degassing during quiescence as a trigger of magma ascent and volcanic eruptions. Scientific reports, 5(1), 1- 7.  
+
+Page 7:  
+
+Line 18: This conceptual model is reasonable, but what is the evidence that there is a crystal- rich or crystal- poor mush? What is the petrological evidence for the percentage of crystals? Is it really \(>50\%\) in one reservoir and \(< 50\%\) in the other?  
+
+Page 8:  
+
+Line 2: You should be clear that you have demonstrated this for a particular volcano during a particular time period and not imply this is a universal process: "the upward transport of magma/gas from the magma chamber toward the surface is a stepwise process in an episodic fashion"  
+
+Page 9: I'm glad to see the discussion of gas (finally), but what are the observations on degassing rate measured on the ground or by satellite? Is it really likely that some events have a high gas proportion and over events a few months later have a low gas proportion? Further, the volume discrepancy might not have anything to do with gas, but could be due to additional reservoirs being tapped that were filled long before the current eruption (that may not have a tilt/trémor signature). A volume difference of 6- 8 is at the high range considered by Rivalta and Segall (2008) but maybe appropriate for this arc volcano? What do the authors think this ratio implies? The question of gas in the magma could be uniquely addressed with this dataset.  
+
+Page 11:  
+
+Line4: This sentence is filled with either assumptions or claims that aren't yet supported in the manuscript such as the existence of a "crystal- rich mush and crystal- poor pool"  
+
+Line 6: Is this claim discussed further somewhere? If so, I missed it: "The duration of each deformation event ( \(\sim 50\) s) is much longer than what is expected for crustal earthquakes of similar size and such a slow deformation"
+
+<--- Page Split --->
+
+The perspective could be improved with adding a paragraph or two about how the lessons here could be applied to specific other volcanoes (if possible). Applying the techniques to other volcanoes is mentioned, but what other volcanoes have a similar eruptive style and might be the best targets to investigate? Also, there are many types of eruptions or plumbing systems for which these types of analysis would not work and should be mentioned as well. If additional space is needed, I suggest dropped Figs. 5 and 6 below.  
+
+The data availability statement is not clear: Are the tilt and seismic waveforms available from the link provided? Also the statement that data products are available by request is no longer considered a best practice (for example, it is not allowed by AGU). These data products are not required to be made available in a public repository, but it would add great value if they were. Does UCL have such a repository?  
+
+Fig.1: What are BCU and BYA chambers? They are not mentioned in the caption. Also, what is the depth of the low velocity zone, and other features listed in the legend (maybe refer to Fig. 3)?  
+
+Fig. 2: Some more details are needed in the caption. How many events are stacked together here? How were events horizontally aligned?  
+
+Fig. 3: Where do the horizontal and vertical displacements in a and b come from? GNSS? Or integrated from seismometer? I could be helpful in d to show the depths of the features shown in a, b, and c: where are the low velocity zone, BYA and BCA chambers, inverted source location (Red Cross) and new Mogi (black circle)? Why is there an aquifer labeled in d? I don't think the aquifer is mentioned in the caption or the text. I'm also confused about what is happening in e, f, and g. What are the red and blue dotted lines at the line labeled LPT? What are the arrow at the SPT line? What are the arrows in- between the LPT and SPT lines? What physical processes do these arrows/features represent?  
+
+Fig. 4a: where does the accumulative net volume change come from? The data used to calculate this should be mentioned in the caption. I do not user stand the labels that say, for example, delta V magma \(< <\) delta V gas during time period 1. It looks like the volume change is basically flat during this time period, so shouldn't these two volume changes be in approximate balance instead of orders of magnitude different (as implied by using \(< <\) )?  
+
+Fig. 4b: What is "outgassing potential"? I haven't heard this term before and the phrase used in the caption is still confusing: "the moment ratio between pressurization and 10 depressurization LPTs"  
+
+Fig. 4d: Where is the volume change rate being measured? SMSZ? The caption says the black crosses are the "inflation event" but I think this should be "inflation events." What is the geodetic data used to estimate the green cross? Is this the same tilt data described in this paper or something else, like GNSS? What is the time period of the geodetic data? In general, it should be noted that these volume change rates are being measured over vastly different time periods and the time periods should be mentioned in the caption.  
+
+Fig. 4e: Considering there are only 2 data points, it does not seem wise to draw a line between them.  
+
+Figures 5 and 6 have minimal discussion in the text and do not seem to factor into the key conclusions mentioned in the abstract - - could they be removed? They probably deserve further discussion in a separate paper. Fig. 5 has a huge amount of information that isn't discussed in the main text. In particular, I think Fig. 6 is confusing and possibly misleading. It seems to take a single volcanic system and wildly extrapolate it to all systems worldwide. This figure seems to imply that the magma plumbing system of Aso is relevant to all types of eruptions from Rhylites to flood basalts which is clearly not true - - we have enough information to know that the plumbing of Aso is not widely applicable to all volcanoes. The authors should consider what is the point this figure is trying to make and if it is already made successfully in the text. If the point is that "Composition, viscosity, rheology and tectonic settings govern the recurrences of episodic
+
+<--- Page Split --->
+
+deformation" then this point can be made adequately in the text without confusing the reader.  
+
+Fig. 6: "providing the glue" is a confusing phrase here  
+
+Reviewer #3 (Remarks to the Author):  
+
+Review comments on Episodic transport of discrete magma batches beneath Aso volcano By Jieming Niu, Teh- Ru Alex Song  
+
+This manuscript is quite interesting and is sufficiently valuable to publish on Nature Communications.  
+
+Major comments  
+
+1) Source of tilt offset is discussed in the relation of crystal rich and crystal poor zones. I cannot well understand how to relate the source of tilt off set with the rate of crystal. The reference 48 investigate volatile from the viewpoint of petrology. If source of tilt offset is related with volatile-rich and volatile-poor zones, this might be better understood.  
+
+2) Source of tilt offset is also discussed on brittle-ductile transition zone. However, the authors assumes that tilt offset is induced by volume change of a combination model of tensile crack and explosive source in elastic medium and this model does not include fracture. Brittle and ductile are manners of fracture. Regarding to the model of elastic deformation, source of tilt offset should be discussed on difference in elastic constants along the magma plumbing system.  
+
+3) The significance of comparison of 2011-2016 eruptivity of Aso with basaltic eruptivity is not well understood. As mentioned in the text, long-term eruption rate of andesitic volcanoes is lower than basaltic volcanoes. I cannot find a significance of comparison of eruption rate between andesitic and basaltic volcanoes in this manuscript. The eruption 2011-2016 is an eruptive activity of Aso, however it does not cover all the eruptivity of Aso. If compared, the eruption 2011-2016 should be compared with past eruptivity of Aso or long-term eruptivity of the volcano.  
+
+Minor comments  
+
+P3L20  
+
+"natural period" -> band width  
+
+P4L12  
+
+LP signal east- west is much weaker, but north- south is stronger than N.ASIV.  
+
+P4L13 "These observations strongly indicate that the source of the tilt offset is spatially separated from the LPT source"  
+
+It is possible, but is it necessary to examine source difference between tilt offset and LPT?  
+
+P4L14 "which is near the active Naka- dake first crater"  
+
+Show references or see Method.  
+
+P5L3 global waveform stacks  
+
+What do you mean "global"? How many LPTs were stacked?  
+
+P5L9 relatively steady  
+
+Almost the same?  
+
+P5L21 Fig. S4  
+
+Fig. S4 is "Synthetic amplitude- distance decay against static and filtered waveforms". Inserting Fig. S4 explains why you choose 100- 200s ULP band?  
+
+P7L2 "SMSZ"  
+
+"SMSZ" firstly appeared here. This should be explained.  
+
+P7L5 intense  
+
+Is this word necessary? What do you mean "intense unrest"?  
+
+P7L18 "a deeper reservoir (i.e., a crystal- rich mush) and temporarily stalled in the SMSZ (i.e., a crystal- poor pool)"  
+
+The reference 48 investigate volatile. How does it relate to crystal? Are there any evidence for crystal- poor in the SMSZ?
+
+<--- Page Split --->
+
+P8L1- L7  
+
+For estimation of volume increase, Method should be referred.  
+
+P8L13- L22  
+
+This hypothesis is true. As the authors show clearly, LPT is a short- lived phenomenon. What does the author mention, comparing short- lived volume change with long- term eruptivity?  
+
+P10L11- L13 "While the estimated mass flow rate in Aso is lower than those estimated in basaltic eruptions by an order of magnitude (Fig. 5), such a difference is consistent with the disparity in the average volcanic output rate between basaltic eruptions and andesitic eruptions"  
+
+In this case, Aso means activity 2011- 2016? Which eruption does "basaltic eruption" indicate? The first half is comparison of short term activity. Second half is long- term comparison. Long- term comparison may be true. But short- term comparison is case- by- case. It is not necessary to be consistent.  
+
+P22L11Takahiro -> Takahiro  
+
+FiguresP24L9 October 8, 2016Need time of onset of the eruption.(b) N.ASHV.LE is tilt of east- side down at station N.ASHV?  
+
+Is Figure 6 needed?
+
+<--- Page Split --->
+
+In the following response, the comments by the reviewers are shown in italic, our responses are shown in bold. The page and line numbers in the original manuscript are noted as P?L??. The page and line numbers In the revised manuscript are noted as nP?L??. The responses discussed in the rebuttal letter are also highlighted in yellow in the revised manuscript.  
+
+Reviewer #1 (Remarks to the Author):  
+
+Overall impression  
+
+The article written by Niu and Song shows an exciting interpretation of tons of tilt and seismic signals that repeatedly occurred at Aso volcano, Japan, with precise analysis procedures and its results. It sheds light on the new concept of the magma plumbing system at the volcano, adding a magma storage zone located between the magma chamber and the conduit. It identifies that this zone's inflation and deflation are associated with episodic magma discharge from below and are indeed relating to the surficial phenomena. After calculating the 1- D conduit flow, it improves the public image of magma ascent dynamics from the magma chamber to the surface. It also provides a future direction of geophysical research to real- time evaluate upcoming eruptions. This article seems to be well- organized, and the methods and results are detailed and very interesting for the scientific community widely. I would recommend their article for publishing in the journal after minor revisions of some comments below I pointed out.  
+
+We thank the reviewer for the positive comments and support for the publication.  
+
+Specific comments  
+
+P3 L13 - The authors use the term "long- period tremor (LPT)" for the objected seismic signal in the whole in this article because they respect the nomenclature that had been established in previous papers. However, this term seems to be a local vernacular or jargon only at Aso volcano. I am also suspect that that signal is not a "tremor?" I suggest using the more widely used term VLP or ultra- long- period (ULP) signals in the volcanological field instead of LPT for readers of this journal from more wide backgrounds.  
+
+We thank the reviewer for the suggestion.  
+
+Indeed, the LPT is associated with event- like signal, and it does not resemble typical tremors. We have replaced LPT with the more widely recognized term VLP in the revised manuscript. We also replaced the term SPT with the more widely recognized term LP to represent the long- period signal above the crack- like conduit in Aso volcano.  
+
+P4 L4 - Were the VLP signals the authors checked from the catalog associated individually with any surficial eruption activities at the crater during the erupted periods?  
+
+We thank the reviewer for the comment.  
+
+Some VLPs are indeed associated with individual surficial eruption during the erupted periods. This corroborates the observation by Ishii et al. (2019, EPS). We briefly noted this in the revised manuscript (nP11L6).  
+
+A paper reported that the occurrence of the VLP seismic signal preceded by few seconds from the onset of Strombolian eruption (Ishii et al., 2019, EPS). What the meaning of a difference of such elapsed time (2 min for Oct. 2016 phreatomagmatic eruption and 2- 3 s for Strombolian explosions), as well as the meaning of a difference with or without eruptions? I think such a time difference may not relate to the patterns of the tilt source.  
+
+We thank the reviewer's comment.  
+
+Motivated by the discussions by Ishii et al. (2019), we have added two paragraphs and highlighted the implications of these differences against the inferred magma ascent velocity
+
+<--- Page Split --->
+
+in the SMSZ (nP11L3- L13). We conceive that the time delay between VLP and the onset of Strombolian eruption can be reasonably attributed to gas ascent in the shallow conduit under slow magma ascent (e.g., 0.01- 0.1 m/s). On the other hand, the 120 s time delay between Event 2 and the onset of 2016 phreatomagmatic eruption can be reconciled with the relatively magma ascent in the SMSZ (e.g., \(\sim 7 \mathrm{m / s}\) ).  
+
+On the other hand, VLP and synchronous inflation/deflation event does occur with or without eruption. It implies that the occurrence of VLP is likely a necessary condition for surface eruption. As discussed by Niu & Song (2020), conduit/plug permeability and overpressure also play a key role in facilitating surface eruptive activities. We have added a short note in the revised manuscript (nP8L20- L24).  
+
+Are there several types of VLPs with similar waveforms, but specific properties are different? Can the author comment on this topic at any part of this article?  
+
+We thank the reviewer's comment.  
+
+We have commented their differences (e.g., resonance period, initial polarity) in the revised manuscript (nP3L10- L14).  
+
+P4 L20 - Can the authors show a result of the same method using another template (Event 2) in the supplementary? I could not find validity to use Event 1 (not 2) as a reference signal in this article. I am convincing that the conclusion will be the same as this article if Event 2 is used.  
+
+We thank the reviewer for the comments.  
+
+Choosing Event 1 as the template has one advantage in that it potentially minimises the interference from the eruption signal at long period (i.e., 50- 250 sec). However, the waveforms from Event 1 and Event 2 are highly correlated, using Event 2 as the reference event does not result in noticeable difference in the detection of inflation or deflation events. The inverted source location and mechanism from Event 2 are practically the same as those inverted from Event 1 (see figure below). We do not include the inversion result from Event 2 in this paper to avoid redundancy. We emphasized the high correlation between Event 1 and Event 2 in the revised manuscript (nP3L24- P4L1).  
+
+![PLACEHOLDER_9_0]
+
+<center>Rebuttal Figure 1 Posterior probability functions of the inverted source parameters from Event 2. Notations are the same as Supplementary Figure 12 in the manuscript. </center>
+
+<--- Page Split --->
+
+However, curious about the time evolution of the VLPs for the case of 2.  
+
+We thank the reviewer's comment.  
+
+Perhaps there is a misunderstanding here. We emphasize that waveforms from Event 1 and Event 2 are practical identical in the period of our interest (i.e., 50- 250 sec). Therefore, VLP signal of 10 sec period is not the focus of our detection. Rather, the timing of VLP was used to detect inflation or deflation events.  
+
+In a separate note, the evolution of VLPs has been thoroughly documented by Niu & Song (2020, JVGR) and we refer the reader to this article for details.  
+
+P5 L15 - The authors should explain the excitation mechanism of the SPT concerning the occurrence of VLP as well as the series of phenomena starting from a deeper place (top of the magma chamber; the authors argue). I could not imagine why the SPT starts at the top portion of the conduit (VLP source) at first, and 10 s later, the VLP occurs (it is the same time as the SPT peak), whenever the events are inflation or deflation.  
+
+We thank the reviewer for this comment.  
+
+Following the discussions by Kaneshima et al. (1996) and Kawakatsu et al. (2000), we have added a paragraph to note the possible processes associated with VLP and LP. We emphasize the stress perturbation induced during the initial stage of the inflation/deflation event results in fractures in the conduit plug above the shallow crack- like conduit, promoting LP and VLP (nP6L7- L12).  
+
+We have appended additional paragraphs in the revised manuscript and elaborated the series of phenomena starting from the deep source near the chamber roof (nP7L3- L28). We caution that this is a tentative interpretation at this point and a more elaborated discussion is beyond the main scope of this manuscript.  
+
+P6 L10 - The authors cited Hata et al. (2018, JGR) in this article; however, it seems inappropriate. At least another paper of Hata et al. (2018, JGR, 10.1029/2018JB015951), or much preferable Matsushima et al. (2020, EPS) showing the revised model of the Hata et al.'s result should be cited.  
+
+We thank the reviewer for pointing out this issue.  
+
+We have cited Hata et al. (2018, JGR, 10.1029/2018JB015951) and Matsushima et al. (2020, EPS) in the revised manuscript (nP6L23).  
+
+P6 L20 - Add any comment on how to make seismicity around the roof of the magma chamber if gas- dominant materials transport upward.  
+
+We thank the reviewer for this comment.  
+
+As also included in the reply earlier, we noted the role of gassed magma in the inflation/deflation events in the revised manuscript (nP7L7- L13). The discussions are in line with the suggestion by the reviewer3 that gas near at the top of the magma chamber may facilitate episodic brittle failure due to increasing magma buoyancy.  
+
+P7 L1 - Can the authors show other evidence of inflation and deflation during these periods? It would be enough to cite some papers of InSAR or GNSS that can strengthen their argument.  
+
+We thank the reviewer for the comment.  
+
+As noted in the original manuscript, the signals associated with the inflation or deflation events are on the order of \(1 \mu \mathrm{m}\) in displacement, which is much smaller than the detection threshold of InSAR and GNSS (i.e., mm to cm). We have emphasized the limitation in the revised manuscript (nP2L22- L28).
+
+<--- Page Split --->
+
+Furthermore, as readily discussed in the Methods sections, to identify the displacement or tilt offset, we remove the mean of the background trend before the LPT arrival. As suggested in the original manuscript, these detected events likely represent episodic transport of discrete magma from the roof of the magma chamber to the storage zone directly above, i.e., the source and sink are probably too close to be deciphered from GNSS or InSAR.  
+
+On the other hand, we note that GNSS displacement from JMA show a notable inflation of the magma chamber in mid- July 2014, May 2015 and July 2016, suggesting magma ascent from a deep reservoir \((> - 10\mathrm{km})\) toward the magma chamber. These episodes coincide with a substantial increase in the number of inflation events and magma transport toward the SMSZ. We have briefly discussed this in the revised manuscript (nP8L12- L14).  
+
+P7 L14 - I could not understand the concept relating to SO2 gas emission. Could the author explain more carefully? The authors describe the prominence of inflation events equivalent to pressurization and lowering SO2 emission. The relation between inflation and pressurization is readily accepted because both are relative changes of the SMSZ condition. However, in my understanding, SO2 emission does not seem to this relative change; but relates to the absolute volume of magma transported to a shallow depth. Is this wrong? I could not see a significant correlation between Fig 4b and 4c.  
+
+We thank the reviewer for this comment.  
+
+In the original manuscript, we refer to the observation that the prominence of inflation (deflation) events coincides with a higher proportion of the pressurization (depressurization) VLP event in the crack like conduit and a lower (higher) \(\mathrm{SO}_2\) emission. We do not mean to suggest that there is causal relationship between \(\mathrm{SO}_2\) emission and the prominence of inflation events. As discussed in Niu & Song (2020), the outgassing potential shown in Fig. 4b in the original manuscript was meant to infer conduit/plug permeability. Therefore, we do not necessarily expect a high correlation between Fig. 4b and Fig. 4c in the original manuscript.  
+
+On the other hand, we do observe a clear correlation between rising \(\mathrm{SO}_2\) emission, the accumulative volume of magma transport in the SMSZ and crater bottom temperature before the 2014 Strombolian eruption. We have replaced Fig 4b with the crater bottom temperature (now Fig. 5b in the revised manuscript) highlight these observations in the revised manuscript (nP8L14- L18).  
+
+P9 L11 - Probably the wrong citation; did not Miyabuchi & Hara (2019, EPS) treat the 2016 phreatomagmatic explosion? I guess Ishii (2018, EPS), Ishii et al. (2018, EPS) and Sato et al. (2018, EPS) are more suitable for the mass of ejected materials.  
+
+We thank the reviewer's comment.  
+
+Miyabushi & Hara (2019, EPS) concerned about the eruptive mass of the 2014 strombulian eruption, which is not a suitable citation. On the other hand, the reference cited in the original manuscript was Miyabuchi et al. (2017, JpGU abstract), which reported the eruptive mass from the field survey. We note the result of field survey by Miyabuchi et al., (2017) has also been cited by Ishii (2018), Ishii et al. (2018) and Sato et al. (2018). To keep the reference list not too long, we have cited Ishii (2018, EPS) in the revised manuscript (nP9L27).  
+
+P11 L 13 - Unclear to me.
+
+<--- Page Split --->
+
+We thank the reviewer for the comment.  
+
+We have rephrased and clarified how one may use the amplitude scaling between VLP and the inflation (or deflation) event derived in recent eruptions to evaluate pre- eruptive volume change in the SMSZ that is otherwise inaccessible for historical eruptions without modern data (nP12L11- L16).  
+
+References - It is not good to cite several papers written in Japanese, which most of the readers cannot probably reach and read. Other accessible papers should be referred to in this article as far as possible.  
+
+We thank the reviewer's comment.  
+
+There are four cited papers written in Japanese. The reference (54,69) in the original manuscript have been removed. However, the results in references (40,49) are essential and there is no alternative literature published in English. Therefore, we retain the two references (40, 49) (reference 46, 45 in the revised manuscript).  
+
+40. Ohkura et al. (2009).  
+
+49. Sudo et al. (2006).  
+
+54. Yokoo & Miyabuchi (2015).  
+
+69. Sakaguchi et al. (2008).  
+
+The author also should discuss the proposed plumbing system with the result of Tsutsui & Sudo (2004, JVGR).  
+
+We thank the reviewer's comment.  
+
+We have cited Tsutsui & Sudo (2004) in the proposed plumbing system in the revised manuscript (nP6L20).  
+
+Fig. 2 - I guess the exact time of the onset in longer signals is quite tricky. How about the reading error?  
+
+We thank the reviewer for this comment.  
+
+The onset of the long- period signal is identified when the velocity exceeds the peak- to- peak amplitude of the background noise (Fig. 2). We estimate the picking error of \(\sim 5\) sec or less.  
+
+P37 L 14 - need a reference for assumption (or evaluation)  
+
+We thank the reviewer for the suggestion.  
+
+There was a typo in the Lame's constant in the original manuscript (P37L14). The correct Lame's constant \(\lambda\) is 1.649 GPa. Following Legrand (2000), we assume \(\nu_{p} = 1500 m / s\) , \(\nu_{s} = 800 m / s\) and \(\rho = 1700 kg / m^{3}\) . We have cited this reference in the supplementary. Since the period of interest is 50 sec and longer, as detailed in the reply to the reviewer 3, the absolute value of the elastic constant does not change the source characteristics or the estimated volume change. We noted this in the revised manuscript (nP5L26- L29).
+
+<--- Page Split --->
+
+Reviewer #2 (Remarks to the Author):  
+
+This paper provides interesting new observations of coupled seismic and ground deformation of repeated magma transport events observed at Aso volcano, Japan that are at least sometimes associated with eruptions. These types of high spatial and temporal resolution observations combining these datasets are rare (to my knowledge) and provide a unique perspective on the timing, location, and volume change of magma movements within a volcano. It would be good to get the perspective of an expert on the Aso system, but as far as I can tell, this paper provides new insight into the plumbing system at this volcano and is possibly applicable (in terms of the conceptual model and technique) to other volcanic systems.  
+
+We thank the reviewer for the positive comments and support for the publication.  
+
+I recommend publication after moderate revision. There are several steps in the analysis that are not well documented (see detailed comments below). Further, the paper is unclear in some locations or the discussion is incomplete (again documented below).  
+
+Page 1:  
+
+Line 10 and Page 1 lines 6- 9: How do we know that Aso has a crystal rich mush and that it is relevant for this study? Maybe the magma batches are coming from a crystal poor reservoir? Maybe the crystal rich mush is deeper?  
+
+We thank the reviewer for the comment, which help improve the interpretation.  
+
+We largely agree with the reviewer's suggestion. In the revised manuscript, the abstract has been modified. We note that the crystal- rich mush domain likely remains at a deeper depth (i.e., \(>10\mathrm{km}\) ) after the most recent Aso- 4 caldera forming eruption (e.g., Ishibashi et al., 2018) (nP6L13- L18).  
+
+Line 14: "individual" should be "an individual"  
+
+We thank the reviewer for the comment.  
+
+It has been removed in the abstract in the revised manuscript.  
+
+Lines 16- 24: I found the following sentences unclear and confusing - - can the authors be more specific? What do you mean by "composition dependent"? Is the composition of each magma batch different? Does the last sentence mean that you can forecast the eruption style, plume height, etc. based on the tilt/seismic data described above? If so provide some more details as to how.  
+
+"whereas their recurrences, potentially composition dependent, are regulated by the brittle- to- ductile transition rheology under low differential stress and high strain rate due to the surge of magma from below, regulating long- term volcanic output rate. The magma ascent velocity, decompression rates, and cumulative magma output deduced from the episodic deformation events before recent eruptions in Aso volcano are compatible with retrospective observations of the eruption style, tephra fallouts, and plume heights, promising real- time evaluation of upcoming eruptions."  
+
+We thank the reviewer's comment.  
+
+First, we clarify the misunderstanding that the composition of each magma batch is different. Rather, "composition- dependent" was referred to the hypothesis that the recurrence interval of magma transport in a more silici magma may occur much less frequently than that in a basaltic or andesitic magma. We have removed this sentence in the revised manuscript to avoid confusion.
+
+<--- Page Split --->
+
+The last sentence in the abstract in the original manuscript does imply the possibility of assessing the eruption size and style through the estimated mass flow rate and magma ascent velocity before the upcoming eruption. While the abstract has been rewritten, we simply note the effect of magma composition on the recurrence interval of episodic deformation events (nP10L12- L14).  
+
+Further, the results shown in Figures 4- 6 aren't really described in the abstract.  
+
+We thank the reviewer for the comment.  
+
+As suggested by the reviewers, we removed Figure 6 in the original manuscript. The results in Figs. 4- 5 (Figs. 5- 7 in the revised manuscript) are honoured in the abstract in the revised manuscript.  
+
+Page 2:  
+
+Line 5: the ambient stress state also matters  
+
+We thank the reviewer for this comment.  
+
+We have modified the sentence accordingly in the revised manuscript (nP2L4).  
+
+Line 16: Also the ambient stress state matters - - seismicity will only occur where the rocks are near to failure. Magma can move aseismically if the stress state is not close to failure.  
+
+We thank the reviewer for this comment.  
+
+We agree with the reviewer and have stressed this point in the revised manuscript (nP2L18- L19).  
+
+Page 3:  
+
+Line 8: "signal" should be "signals"  
+
+We thank the reviewer for this comment.  
+
+We have modified the phrase accordingly (nP3L4).  
+
+Line 10: How do we know this is a "shallow hydrothermal reservoir"?  
+
+We thank the reviewer for the comment.  
+
+The presence of a shallow hydrothermal reservoir was inferred from a high electrical conductivity channel in several literatures (Hase et al., 2005; Hata et al., 2016; 2018; Kanda et al., 2008, 2019). We have cited relevant references in the revised manuscript (nP3L7- L9).  
+
+Line 11: use "on" instead of "against"  
+
+We thank the reviewer for this comment.  
+
+This phrase has been modified accordingly (nP3L7).  
+
+Line 12 (and Page 1 line 12): Mentioning the source is near sea level is confusing, how far is this below the surface? It would be better to tell us the depth of the source beneath the surface (or at least tell us both pieces of information).  
+
+We thank the reviewer for the comment.  
+
+We have appended the source depth in the revised manuscript (nP3L9).  
+
+Line 16: Need some introduction to the eruptive cycle - - why was the time period 2011- 2016 chosen? Why not a longer time period?  
+
+We thank the reviewer for this comment.  
+
+The typical eruption cycle in Aso volcano has been summarized previously (e.g., Sudo et al., 2006, Kawakatsu et al., 2000). We have added some details in the introduction (nP3L15- L20).
+
+<--- Page Split --->
+
+Analysis of a larger dataset including activities after 2016 is currently undertaking and will be reported in the future.  
+
+Is the LPT only seen in this time period?  
+
+We thank the reviewer's comment.  
+
+As stated in the original manuscript, LPT can be observed regardless of surface activity and it has also been observed in the past (e.g., Sassa, 1933; Kaneshima et al., 1996; Kawakatsu et al., 2000; Niu & Song, 2020). As noted above, the analysis includes a relatively complete Aso eruption cycle in 2011- 2016.  
+
+Is this the only time period when the patterns described below occur? Or some other reason? As noted above, the analysis and detection of inflation/deflation events includes a relatively complete Aso eruption cycle in 2011- 2016. Analysis of a larger dataset beyond 2016 is currently undertaking and will be reported in the future.  
+
+Line 17: "waveform" should be "waveforms" We thank the reviewer for this comment. We have modified accordingly in the revised manuscript (nP3L23).  
+
+Line 21: How do you know these LPT events are "anomalous"? Where do you define normal or background LPT activity?  
+
+We thank the reviewer for this comment.  
+
+The LPT catalog has been detailed in Niu & Song (2020, JVGR). The amplitude of the two LPT events prior to the 2016 eruption is at least 2 orders magnitude larger than any LPTs we identified in 2011- 2016. We add a brief note in the revised manuscript (nP3L28- nP4L1).  
+
+Page 4:  
+
+Line 4: where do the displacement waveforms come from? Integration of the seismograms? If so, how?  
+
+We thank the reviewer for the comment.  
+
+The vertical and horizontal displacement are integrated from the seismograms. This signal processing has been described in Methods in the original manuscript and in the revised manuscript.  
+
+Line 5: are these events associated with eruptions?  
+
+We thank the reviewer for this comment.  
+
+As shown in Fig S1, these events are not necessarily associated with eruptions. We have noted this in the revised manuscript (nP4L16).  
+
+line 6: What does "east- down" mean? Does that mean tilt toward the east?  
+
+We thank the reviewer for the comment.  
+
+Yes, "east- down" tilt means that tilt toward the east. We have rephrased the sentence in the revised manuscript (nP4L13- L14).  
+
+Line 8- 9: This phrase could be more precise: "between the signal of LPT and the tilt offset". Perhaps: "between the LPT signal and the tilt offset at different stations and in the different components at the same station."  
+
+We thank the reviewer for the suggestion.  
+
+We have rephrased accordingly in the revised manuscript (nP4L17- L21).
+
+<--- Page Split --->
+
+Line 17: Are all the LPT events associated with eruptions?  
+
+We thank the reviewer for this comment.  
+
+As detailed in P3L16 in the original manuscript, LPT is repetitive regardless of surface activity. Hence, they are not necessarily associated with eruptions.  
+
+Are there LPT events that aren't found by the matched filter? If so, what type of events do they represent?  
+
+The detection capability of the matched filter has been detailed in Niu & Song (2020, JVGR) and they have shown a nominal missed pick rate \(\sim 0.1\%\) . When the signal- to- noise ratio is low, the matched filter will not be able to detect or characterize very small LPT events.  
+
+The sections entitled "LPT and synchronous tilt/displacement offset" and "Discovery of the inflation/deflation event beneath Aso volcano" could be better organized. It seems to organized in a chronological manner instead of a logical description of what was discovered - - the first paragraph talks about the 2016 eruption, the next paragraph is about a manual search and the next paragraph is about a matched filter. Instead, why not just discuss the procedure (pointing to the Materials and Methods as needed) and then describe what you found? Maybe organize: this is what we analyzed, describe the 2016 events (including the variation in signals between stations) and then the global stack.  
+
+We thank the reviewer for this comment.  
+
+We have reorganized this section in the revised manuscript (nP3L22- P5L21). Fig. S4 in the original manuscript has now been moved to Fig. S1 in the revised manuscript.  
+
+Page 5,  
+
+line 4: How many events are in the stack?  
+
+We thank the reviewer for this missing information.  
+
+There are 671 and 967 events for the global inflation and deflation stacks, respectively. This is appended in the caption of the revised manuscript.  
+
+Do the number of events vary in time in a systematic manner?  
+
+The number of events is relatively low in 2011- 2013. The activity increases notably in early 2014 and substantially since July 2014. The monthly event numbers are included in the Fig. S5- 10.  
+
+What do the unstacked events look like?  
+
+Waveforms of the unstacked events have been shown in Fig. S1 in the original manuscript and in the revised manuscript.  
+
+Line 5: what does volcanic unrest mean here?  
+
+We thank the reviewer for this comment.  
+
+The volcanic unrest includes surface volcanic activities such as the dried- up of the crater- late, minor phreatic and ash eruptions and incandescent phenomena. We have appended a sentence in the revised manuscript (nP4L27- L28).  
+
+Are there eruptions in 2011- 2014 or are these events occurring without eruption? If so, that is strange - - why do these similar events occur sometimes with eruption and sometimes without? As noted in Fig. 4, there are isolated minor phreatic eruption and ash eruptions in 2011- 2014, but the detected events are generally not associated with eruptions. As noted by Niu & Song (2020), surface eruptions are not only dictated by the overpressure (i.e., magma
+
+<--- Page Split --->
+
+supply), but also the strength/permeability of the conduit plug. We have appended a brief note in the revised manuscript (nP8L20- L24).  
+
+Page 6:  
+
+Line 4: Make clear from the source what your forward model is in terms of source characteristics and elastic structure.  
+
+We thank the reviewer for this comment.  
+
+While the details of the forward model are described in Method of the initial submission, we add a sentence in the revised manuscript to clarify the source characteristics and elastic structure (nP5L26- L27).  
+
+Line 15: What is the composition of the magma batch based on the eruption? One hypothesis is that this portion of magma is ascending because it has accumulated enough gas to become buoyant, so what is known about the gas compositions in the eruptions?  
+
+We thank the reviewer for the comment.  
+
+The composition of the magma batch is basalt- andesitic, and the detailed analysis has been performed by Saito et al. (2018). They estimated that the buoyancy associated with the gassed magma is up to \(\sim 150 - 350 \mathrm{g / cm^3}\) , facilitating magma ascent. We have noted the role of magma buoyancy in the revised manuscript (nP7L7- L12).  
+
+Line 16: How do we know that this is a crystal rich magma? Is this just a guess or is there some evidence from petrology about the source region?  
+
+We thank the reviewer for this comment.  
+
+We refer to earlier response in the rebuttal letter. We also briefly note the crystal- rich mush region in the revised manuscript (nP6L13- L18).  
+
+Line 20: An alternative is that the magma is that there is no new injection of magma and it is cooling/crystallizing, accumulating gas at the top of the reservoir and then episodically having sufficient buoyancy to cause brittle failure. (This is a top- down instead of a bottom- up trigger for eruptions, see for example Girona, T., Costa, F., & Schubert, G. (2015). Degassing during quiescence as a trigger of magma ascent and volcanic eruptions. Scientific reports, 5(1), 1- 7.  
+
+We thank the reviewer for the comment.  
+
+We agree that accumulating gas at the top of the reservoir may facilitate episodic brittle failure and we have included this mechanism in the revised manuscript (nP7L7- L19).  
+
+As noted in the rebuttal letter earlier, GNSS data reported by JMA indicate a slowdown of deflation in the magma chamber in July 2014, May 2015 and July 2016, suggesting a notable injection of magma from a deeper reservoir below toward the bottom of the magma chamber. These episodes correspond to increasing activities of inflation events and rising crater bottom temperature and SO2 emission (Fig. 5). The discussions are added in the revised manuscript (nP8L12- L18).  
+
+Page 7:  
+
+Line 18: This conceptual model is reasonable, but what is the evidence that there is a crystal- rich or crystal- poor mush? What is the petrological evidence for the percentage of crystals? Is it really \(>50\%\) in one reservoir and \(< 50\%\) in the other?  
+
+We thank the reviewer for this comment.  
+
+We note that the crystal- rich mush domain likely remains at a deeper depth (i.e., \(>10 \mathrm{km}\) ) after the most recent Aso- 4 caldera forming eruption (Ishibashi et al., 2018). On the other hand, we have revisited the magma plumbing system recently summarised by Kawaguchi et al. (2021) and modified the interpretation in the revised manuscript. Specifically, we noted
+
+<--- Page Split --->
+
+the process of magma mixing for post- caldera volcanisms beneath Aso volcano (Miyoshi et al. 2011; Miyoshi et al., 2012; Kawaguchi et al., 2021) and inferred discrete magma transport between a chamber with volatile- poor silicic magma and a storage zone of mixed- magma (i.e., SMSZ) (nP6L13- L19).  
+
+Page 8:  
+
+Line 2: You should be clear that you have demonstrated this for a particular volcano during a particular time period and not imply this is a universal process: "the upward transport of magma/gas from the magma chamber toward the surface is a stepwise process in an episodic fashion"  
+
+We thank the reviewer's comment. We have rephrased the sentence in the revised manuscript (nP8L20- L22).  
+
+Page 9: I'm glad to see the discussion of gas (finally), but what are the observations on degassing rate measured on the ground or by satellite?  
+
+We thank the reviewer's comment.  
+
+The observations of \(\mathrm{SO}_2\) emission were from the campaign ground- based sensor and it is noted in the Fig. 5 caption in the revised manuscript.  
+
+Is it really likely that some events have a high gas proportion and over events a few months later have a low gas proportion?  
+
+We thank the reviewer's comment.  
+
+As detailed later in the rebuttal letter, we stress that the tilt- offset in some of the deflation events during the eruption may decay beyond the time scale of our analysis (i.e., \(\sim 1\) hr) and the volume change of these events is likely associated with the transport of gas, rather than magma. We suspect that such deflation events are likely of a small magnitude and lower signal- to- noise ratio.  
+
+We have appended the net volume change estimated from events with a high signal- to- noise ratio (i.e., subset I+II) (Fig. 5a). While the estimated net volume change before the eruption remains the same, the net volume change in the SMSZ during the eruption is more compatible with the net volume change before the eruption. We have expanded and clarified the discussions in the revised manuscript (nP9L11- L24).  
+
+Further, the volume discrepancy might not have anything to do with gas, but could be due to additional reservoirs being tapped that were filled long before the current eruption (that may not have a tilt/tremor signature). A volume difference of 6- 8 is at the high range considered by Rivalta and Segall (2008) but maybe appropriate for this arc volcano?  
+
+What do the authors think this ratio implies? The question of gas in the magma could be uniquely addressed with this dataset.  
+
+We thank the reviewer's comment.  
+
+First, we followed Rivalta & Segall (2008) and included the effect of magma compressibility, on the estimate of volume change in SMSZ (Fig. 6a in the revised manuscript). The details are included in the Methods section. In short, we obtain \(R_{\nu} = 1.1 - 2.3\) , which means that the eruption output (i.e., mass of tephra fallout) is 1.1- 2.3 times of the estimated volume change in the SMSZ. As discussed by Rivalta & Segall (2008), \(R_{\nu}\) is always greater than 1. However, in the original manuscript, we have shown that the estimated volume change in the SMSZ is 6- 8 times of the DRE of tephra fallout, or equivalently \(R_{\nu} \sim 0.12 - 0.14 < 1\) . Therefore, we think the volume difference is not a result of magma compressibility.
+
+<--- Page Split --->
+
+We again refer to the discussion in the revised manuscript (nP9L11- L24).  
+
+Page 11:  
+
+Line4: This sentence is filled with either assumptions or claims that aren't yet supported in the manuscript such as the existence of a "crystal- rich mush and crystal- poor pool"  
+
+We thank the reviewer for the comment.  
+
+As noted earlier in the rebuttal letter, we have revisited the magma plumbing system recently summarised by Kawaguchi et al. (2021) and slightly modified the interpretation in the revised manuscript (nP6L13- L18). Specifically, we noted the process of magma mixing for post- caldera volcanisms beneath Aso (Miyoshi et al. 2011; Miyoshi et al., 2012; Kawaguchi et al., 2021) and inferred discrete magma transport between a chamber with volatile- poor silicic magma and a storage zone of mixed- magma (i.e., SMSZ).  
+
+On the other hand, we note that the crystal- rich mush domain likely remains at a deeper depth (i.e., \(>10\mathrm{km}\) ) after the most recent Aso- 4 caldera forming eruption (Ishibashi et al., 2018). This is briefly noted in the revised manuscript (nP6L13- L19). We also append a short paragraph in the introduction to lay out the background on the transition of magma plumbing system from caldera- forming eruptions to post- caldera volcanisms (nP2L7- L13).  
+
+Line 6: Is this claim discussed further somewhere? If so, I missed it: "The duration of each deformation event ( \(\sim 50\) s) is much longer than what is expected for crustal earthquakes of similar size and such a slow deformation"  
+
+We thank the reviewer for this comment.  
+
+To put the discussion in a proper context, we have rearranged the paragraph in the revised manuscript (nP7L24- L28).  
+
+The perspective could be improved with adding a paragraph or two about how the lessons here could be applied to specific other volcanoes (if possible). Applying the techniques to other volcanoes is mentioned, but what other volcanoes have a similar eruptive style and might be the best targets to investigate? Also, there are many types of eruptions or plumbing systems for which these types of analysis would not work and should be mentioned as well. If additional space is needed, I suggest dropped Figs. 5 and 6 below.  
+
+We thank the reviewer's comment.  
+
+We have added a paragraph to discuss the applicability of the detection and identified volcanoes where the inference may be possible (nP12L17- L24). Following the suggestion by the review 2 and reviewer 3, we have removed Fig. 6 in the original manuscript.  
+
+The data availability statement is not clear: Are the tilt and seismic waveforms available from the link provided? Also the statement that data products are available by request is no longer considered a best practice (for example, it is not allowed by AGU). These data products are not required to be made available in a public repository, but it would add great value if they were. Does UCL have such a repository?  
+
+We thank the reviewer for the comment.  
+
+The data availability and code availability have been restructured according to Nature Communications requirement. The broadband and tilt waveforms are available from the provided link. The catalogue and code can be obtained upon request.  
+
+Fig.1: What are BCU and BYA chambers? They are not mentioned in the caption. Also, what is the depth of the low velocity zone, and other features listed in the legend (maybe refer to Fig. 3)? We thank the reviewer for this comment.
+
+<--- Page Split --->
+
+We refer the depths info to Fig. 3c.  
+
+Fig. 2: Some more details are needed in the caption. How many events are stacked together here? We thank the reviewer for the comment. We have revised the caption. The number of events used for the inflation and deflation stacks is 671 and 967, respectively.  
+
+How were events horizontally aligned? We clarified in the caption that the traces are aligned with respect to the onset of LP.  
+
+Fig. 3: Where do the horizontal and vertical displacements in a and b come from? GNSS? Or integrated from seismometer?  
+
+We thank the reviewer for this comment.  
+
+The horizontal and vertical displacement are obtained from the broadband seismograms and the data processing has been detailed in the Method section (Estimate the static displacement offset from broadband seismograms).  
+
+I could be helpful in d to show the depths of the features shown in a, b, and c: where are the low velocity zone, BYA and BCA chambers, inverted source location (Red Cross) and new Mogi (black circle)?  
+
+We thank the reviewer's comment. We show the depths of all features in Fig. 3c with a legend.  
+
+Why is there an aquifer labeled in d? I don't think the aquifer is mentioned in the caption or the text.  
+
+The aquifer near the crack- like conduit has been widely discussed in the literature. We have noted the aquifer in the revised manuscript (nP6L6- L12)  
+
+I'm also confused about what is happening in e, f, and g. What are the red and blue dotted lines at the line labeled LPT? What are the arrow at the SPT line? What are the arrows in- between the LPT and SPT lines? What physical processes do these arrows/features represent?  
+
+We thank the reviewer's comment.  
+
+We have modified the caption and append the details (Fig 4 in the revised manuscript).  
+
+Fig. 4a: where does the accumulative net volume change come from? The data used to calculate this should be mentioned in the caption.  
+
+We thank the reviewer for the comment.  
+
+As already noted in the original manuscript (P8L8- L12), the accumulative net volume change is calculated from the monthly volume change associated with the monthly inflation/deflation event stacks. We also mentioned this information in the revised manuscript (nP9L1- L5).  
+
+I do not user stand the labels that say, for example, delta \(V\) magma \(< <\) delta \(V\) gas during time period 1. It looks like the volume change is basically flat during this time period, so shouldn't these two volume changes be in approximate balance instead of orders of magnitude different (as implied by using \(< <\) )?  
+
+We thank the reviewer for this suggestion.  
+
+The volume change during episode 1 is not flat. As noted in earlier reply, we have modified the text to clarify the discussion in the revised manuscript (nP9L11- L17). In particular, we have appended the net volume change from events with a high signal- to- noise ratio (i.e., subset I+II) (Fig. 5a in the revised manuscript) to facilitate the discussion.
+
+<--- Page Split --->
+
+Fig. 4b: What is "outgassing potential"? I haven't heard this term before and the phrase used in the caption is still confusing: "the moment ratio between pressurization and 10 depressurization LPTs"  
+
+We thank the reviewer for the comment.  
+
+To avoid confusion and focus on the observations, we replace Fig. 5b with crater bottom (wall) temperature and append discussions in the revised manuscript (nP8L14- L18).  
+
+Fig. 4d: Where is the volume change rate being measured? SMSZ? The caption says the black crosses are the "inflation event" but I think this should be "inflation events.  
+
+We thank the reviewer for this comment.  
+
+Yes, the volume change rates refer to the SMSZ. We have used circle and cross to indicate the averaged rate and single- event rate, respectively. Fig. 4d has now been rearranged as Fig. 6b in the revised manuscript.  
+
+" What is the geodetic data used to estimate the green cross? Is this the same tilt data described in this paper or something else, like GNSS? What is the time period of the geodetic data? The geodetic data corresponding to the green cross is levelling data. The time period of the geodetic data is from 1958 to 2004. This has been clarified in the revised manuscript (Fig. 6b).  
+
+In general, it should be noted that these volume change rates are being measured over vastly different time periods and the time periods should be mentioned in the caption.  
+
+We thank the comment by the reviewer.  
+
+We have modified the figure and the caption to highlight the difference. In addition, we use a different symbol to highlight the volume change rate for a single event (i.e., time scale of \(\sim\) 100 s).  
+
+Fig. 4e: Considering there are only 2 data points, it does not seem wise to draw a line between them.  
+
+The line is not a fitting line. It simply indicates a 1- to- 1 relationship between the estimated mass change in SMSZ and the mass of tephra fallout. If the estimated mass change in SMSZ is equal to the mass of tephra fallout, the data point will fall on the line.  
+
+Figures 5 and 6 have minimal discussion in the text and do not seem to factor into the key conclusions mentioned in the abstract - - could they be removed? They probably deserve further discussion in a separate paper. Fig. 5 has a huge amount of information that isn't discussed in the main text.  
+
+We thank the review for the suggestion.  
+
+We have removed Fig. 6 in the original manuscript. On the other hand, we have expanded the discussions for Fig. 5 (now Fig. 7) in the revised manuscript (nP11L26- nP12L5).  
+
+In particular, I think Fig. 6 is confusing and possibly misleading. It seems to take a single volcanic system and wildly extrapolate it to all systems worldwide. This figure seems to imply that the magma plumbing system of Aso is relevant to all types of eruptions from Rhyolites to flood basalts which is clearly not true - - we have enough information to know that the plumbing of Aso is not widely applicable to all volcanoes. The authors should consider what is the point this figure is trying to make and if it is already made successfully in the text. If the point is that "Composition, viscosity, rheology and tectonic settings govern the recurrences of episodic deformation" then this point can be made adequately in the text without confusing the reader.
+
+<--- Page Split --->
+
+Following the suggestion by the reviewer 3 and another reviewer, we remove Fig. 6 in the original manuscript since the main point has been discussed in the text. Instead, we add a sentence in the revised manuscript to emphasize that magma composition, viscosity, rheology and tectonic settings govern the recurrences of episodic deformation (nP10L12- L14).  
+
+Fig. 6: "providing the glue" is a confusing phrase herePlease see above reply and Fig. 6 in the original manuscript has been removed.
+
+<--- Page Split --->
+
+Reviewer #3 (Remarks to the Author):  
+
+Review comments on Episodic transport of discrete magma batches beneath Aso volcano By Jieming Niu, Teh- Ru Alex Song  
+
+This manuscript is quite interesting and is sufficiently valuable to publish on Nature Communications.  
+
+Major comments  
+
+1) Source of tilt offset is discussed in the relation of crystal rich and crystal poor zones. I cannot well understand how to relate the source of tilt off set with the rate of crystal. The reference 48 investigate volatile from the viewpoint of petrology. If source of tilt offset is related with volatile-rich and volatile-poor zones, this might be better understood.  
+
+We thank the reviewer for the comments. As pointed out by the reviewer, the reference 48 (Kawaguchi et al., 2021) did not directly address crystal-rich or crystal-poor zone in the magmatic system beneath Aso caldera.  
+
+As noted earlier in the rebuttal letter, following latest petrological studies (e.g., Kawaguchi et al., 2021), we note magma mixing between volatile-poor silicic magma at the magma chamber ( \(\sim 4 - 10 \mathrm{km}\) ) and volatile-rich basaltic magma coming from a deeper reservoir ( \(>10 \mathrm{km}\) ). Furthermore, the storage depth of mixed-magma is determined at \(\sim 2 - 4 \mathrm{km}\) depth (or \(\sim 1 - 3 \mathrm{km}\) BSL). The source of tilt is interpreted as a transport of magma batch between the top of magma chamber and the storage zone of mixed magma (SMSZ discussed in the manuscript). We have modified the interpretation in the revised manuscript (nP6L13- L18, nP6L23- L28). The transport process is also elaborated in the revised manuscript (nP7L3- L28).  
+
+2) Source of tilt offset is also discussed on brittle-ductile transition zone. However, the authors assumes that tilt offset is induced by volume change of a combination model of tensile crack and explosive source in elastic medium and this model does not include fracture. Brittle and ductile are manners of fracture.  
+
+We thank the comment by the reviewer.  
+
+Here we disagree with the reviewer's comment. As noted by Aki & Richard (2002, Quantitative Seismology, chapter 3) and many other seismology textbooks, seismic moment tensor is a general force- equivalent representation of internal sources, including fracture, explosion and tensile- crack inside the earth. On the other hand, to reiterate, our seismic moment tensor inversion shows that the source of the tilt- offset has a predominant volumetric component and \((\sim 80\%)\) and a minor normal- fault component \((\sim 20\%)\) .  
+
+Regarding to the model of elastic deformation, source of tilt offset should be discussed on difference in elastic constants along the magma plumbing system.  
+
+As noted by Aki & Richard (2002, Quantitative Seismology, chapter 3), the elastic moduli used in the force- equivalent representation of internal sources are constants appropriate for the wall rock (or unaltered rock). While it is likely that the elastic constant within the magma plumbing system may vary, it does not change seismic wave excitation in the frequency band of our interest (i.e., \(>50\) sec). We have briefly discussed the effect of elastic constant on source inversion and the volume change in the revised manuscript (nP5L26- L29).  
+
+The inverted source location and geometry only depends on the displacement or tilt ratio among different stations or/and different components. We also inverted the source with a different velocity structure and the effect of elastic constant on the solution is minimal.
+
+<--- Page Split --->
+
+3) The significance of comparison of 2011-2016 eruptivity of Aso with basaltic eruptivity is not well understood. As mentioned in the text, long-term eruption rate of andesitic volcanoes is lower than basaltic volcanoes. I cannot find a significance of comparison of eruption rate between andesitic and basaltic volcanoes in this manuscript.  
+
+We thank the reviewer for the comment.  
+
+We have added a paragraph and noted the relevance of the comparison in the revised manuscript (nP11L26-nP12L5). Comparing mass discharge rate between basaltic and basalt-andesitic volcanoes help elucidate the regime where the mass discharge rate can be approximated by the mass flow rate.  
+
+The eruption 2011- 2016 is an eruptive activity of Aso, however it does not cover all the eruptivity of Aso. If compared, the eruption 2011- 2016 should be compared with past eruptivity of Aso or long- term eruptivity of the volcano.  
+
+We largely agree with the reviewer's comment. In Fig. 4d in the original manuscript, we did compare the averaged eruption output rate in 2011- 2016 against the output rate in historical eruptions and long- term eruption output rate over the geological time scale. These comparisons have been noted in P8L8- L18 in the original manuscript. In the revised manuscript, we also appended the magma discharge rates of the 1979 and 1989 eruptions in the Fig. 6 for comparisons. We have modified the text accordingly (nP11L16- L20).  
+
+Minor comments P3L20  
+
+"natural period" -> band width  
+
+The use of natural period was correct. It is used to describe the resonant period of a seismometer (pp.175, Lay and Wallace, 1995).  
+
+P4L12  
+
+LP signal east- west is much weaker, but north- south is stronger than N.ASIV.  
+
+We thank the reviewer for the comment. The change has been made in the revised manuscript (nP4L17- L21).  
+
+P4L13 "These observations strongly indicate that the source of the tilt offset is spatially separated from the LPT source"  
+
+It is possible, but is it necessary to examine source difference between tilt offset and LPT?  
+
+We thank the reviewer for the comment.  
+
+As discussed in P4L8- L13 in the original manuscript, the difference in the amplitude between LPT and tilt offset among different seismic station or/and channels is a direct indication that the source of tilt offset must differ from the source of LPT, either in location or mechanism. We have rephrased the sentence to clarify this conjecture (nP4L23- L25).  
+
+P4L14 "which is near the active Naka- dake first crater"  
+
+Show references or see Method.  
+
+We thank the reviewer's comment and have cited relevant references in the revised manuscript (nP4L25).  
+
+P5L3 global waveform stacks  
+
+What do you mean "global"? How many LPTs were stacked?  
+
+We thank the reviewer for the comment. The term "global" is used to highlight waveform stacks over the period of unrest (2011- 2014 August) and Strombolian eruption (Nov. 2014-
+
+<--- Page Split --->
+
+Apr 2015). The number of LPT events used for stacking is shown in Fig. 2 and noted in the caption.  
+
+P5L9 relatively steady Almost the same?  
+
+We thank the reviewer's comment. We have rephrased the sentence in the revised manuscript (nP5L4).  
+
+P5L21 Fig. S4  
+
+Fig. S4 is "Synthetic amplitude- distance decay against static and filtered waveforms". Inserting Fig. S4 explains why you choose 100- 200s ULP band?  
+
+We thank the reviewer for the comment.  
+
+We have modified the sentences accordingly in the revised manuscript (nP4L3- L6).  
+
+P7L2 "SMSZ"  
+
+"SMSZ" firstly appeared here. This should be explained.  
+
+We thank the reviewer for the comment. The SMSZ is first defined and noted in the revised manuscript (nP6L21).  
+
+P7L5 intense  
+
+Is this word necessary? What do you mean "intense unrest"?  
+
+We thank the reviewer for the comment.  
+
+We have added a brief note to define specific activities in the revised manuscript (nP8L2- L4).  
+
+P7L18 "a deeper reservoir (i.e., a crystal- rich mush) and temporarily stalled in the SMSZ (i.e., a crystal- poor pool)"  
+
+The reference 48 investigate volatile. How does it relate to crystal? Are there any evidence for crystal- poor in the SMSZ?  
+
+We thank the reviewer for the comment.  
+
+As noted earlier in the rebuttal letter, we have modified the interpretation and the discussions in the revised manuscript (nP6L13- L19).  
+
+P8L1- L7  
+
+For estimation of volume increase, Method should be referred.  
+
+We thank the reviewer for the comment. This statement has been referred to the Method in the revised manuscript (nP9L1- L5).  
+
+P8L13- L22  
+
+This hypothesis is true. As the authors show clearly, LPT is a short- lived phenomenon. What does the author mention, comparing short- lived volume change with long- term eruptivity?  
+
+We thank the reviewer for the comment.  
+
+As noted in the original manuscript LPT (termed VLP in the revised manuscript) has been observed over multiple eruption cycles since 1920s (Sassa, 1935; Sakaguchi et al., 2008). This provides a means to detect synchronous VLP and inflation/deflation event and evaluate single- event volume change for historical eruptions. This point has been addressed in the revised manuscript (nP12L11- L16).  
+
+P10L11- L13 "While the estimated mass flow rate in Aso is lower than those estimated in basaltic eruptions by an order of magnitude (Fig. 5), such a difference is consistent with the disparity in the average volcanic output rate between basaltic eruptions and andesitic eruptions"
+
+<--- Page Split --->
+
+In this case, Aso means activity 2011- 2016? Which eruption does "basaltic eruption" indicate?  
+
+We thank the reviewer for the comment.  
+
+Yes, the mass discharge rate was referred to Aso activity 2011- 2016. In the original manuscript, the basaltic eruptions refer to data shown in Fig 5 (solid diamonds in Fig. 7 in the revised manuscript) where both the decompression rate and the mass discharge rate are available (Barth et al., 2019). On the other hand, the relevant discussion has been revised in the revised manuscript (nP11L26- nP12L5).  
+
+The first half is comparison of short term activity. Second half is long- term comparison. Long- term comparison may be true. But short- term comparison is case- by- case. It is not necessary to be consistent.  
+
+As noted above in the rebuttal letter, the relevant discussion can be referred to the revised manuscript (nP11L26- nP12L5).  
+
+P22L11Takahiro -> TakahiroWe thank the reviewer for the comment. This has been modified accordingly.  
+
+FiguresP24L9 October 8, 2016Need time of onset of the eruption.The onset of the eruption is marked by the red line. We have modified the caption of Fig. 1b accordingly.  
+
+(b) N.ASHV.LE is tilt of east-side down at station N.ASHV?Yes, we have also clarified this in the caption.  
+
+Is Figure 6 needed?We thank the reviewer for the comment.As suggested by the reviewers, this figure has been removed.
+
+<--- Page Split --->
+
+## REVIEWERS' COMMENTS  
+
+Reviewer #2 (Remarks to the Author):  
+
+I thank the authors for addressing my comments in the first round of review. I think they have adequately modified their manuscript and I recommend publication in its present form.  
+
+Reviewer #3 (Remarks to the Author):  
+
+I think the manuscript is well revised according to the comments by the reviewers. This manuscript documents magma plumbing system of Aso volcano from the deep chamber to the shallow conduit and reveals SMSZ to connect the two magmas. The two stages; pre- eruptive and Strombolian eruption are well separated by extracting minor offset of the tilt and seismic events. This attains at a level to publish on nature communications, supporting high- precision data.  
+
+The followings may be mistyped. P2 L27 sthe- >the Fig1 L6 NewMogi - > New Mogi Fig2 L8 acasual - > casual
+
+<--- Page Split --->
+
+In the following response, the comments by the reviewers are shown in italic, our responses are shown in bold. The page and line numbers in the original manuscript are noted as P?L??. The responses discussed in the rebuttal letter are also highlighted in yellow in the revised manuscript.  
+
+Reviewer #2 (Remarks to the Author):  
+
+I thank the authors for addressing my comments in the first round of review. I think they have adequately modified their manuscript and I recommend publication in its present form. We thank the reviewer for the positive comments and support for the publication.  
+
+Reviewer #3 (Remarks to the Author):  
+
+I think the manuscript is well revised according to the comments by the reviewers. This manuscript documents magma plumbing system of Aso volcano from the deep chamber to the shallow conduit and reveals SMSZ to connect the two magmas. The two stages; pre-eruptive and Strombolian eruption are well separated by extracting minor offset of the tilt and seismic events. This attains at a level to publish on nature communications, supporting high- precision data. We thank the reviewer for the positive comments and support for the publication.  
+
+The followings may be mistyped.  
+
+P2 L27 sthe- >the This has been corrected as P2L27 in the revised manuscript.  
+
+Fig1 L6 NewMogi - > New Mogi This has been corrected as P30L8 in the revised manuscript. "NewMogi" in the figure has also been replaced by "New Mogi".  
+
+Fig2 L8 acasual - > casual This has been corrected as P31L8 in the revised manuscript.  
+
+Note we slightly revise Fig. 5b and append the temperature data from ground- based thermal camera (Cigolini et al., 2018), filling the data gap between late 2014 and 2016. We append two references summarizing latest efforts on deep low- frequency earthquakes (P12L25). We slightly refine and make a more precise description of source mechanism (P6L3- 5) and append uncertainty estimates in the supplementary Table 2.  
+
+Finally, we have also followed the checklist and slightly reformat the manuscript. As suggested by the editorial team, we include an image to be featured in Nature Communications. We believe the image vividly illustrates the background crater- lake and diverse eruption styles in Aso volcano. The original images are credited to Dr. Akihiko Yokoo in Aso Volcano Observatory.
+
+<--- Page Split --->

peer_reviews/129058fcec31e99c0819f8921d0245dbb3bd7403d48f93401d336635a4c8ebb2/supplementary_0_Peer Review File/images_list.json

@@ -0,0 +1 @@
+[]

peer_reviews/129058fcec31e99c0819f8921d0245dbb3bd7403d48f93401d336635a4c8ebb2/supplementary_0_Peer Review File/supplementary_0_Peer Review File.mmd

@@ -0,0 +1,167 @@
+
+# nature portfolio  
+
+Peer Review File  
+
+Activator- blocker model of transcriptional regulation by pioneer- like factors  
+
+![PLACEHOLDER_0_0]
+  
+
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+
+## REVIEWER COMMENTS  
+
+Reviewer #1 (Remarks to the Author):  
+
+In this manuscript, Riesle and Gao et al. analyzed ATAC- seq, H3K27ac ChIP- seq, and transcriptome of zebrafish mutants of ZGA regulator TFs. They found the existence of synergistic and antagonistic enhancer types, and Pou5f3 and Nanog can function as activator or blocker in antagonistic enhancers. The topic is unquestionably important for the field, but this reviewer feels that the novelty of the study seems weak. I have several major concerns as described below.  
+
+Regarding the novelty, the fact that the pioneer factors (Pou5f3 and Sox19b) is preventing premature expression for some of the genes during ZGA has already been reported by the same group (Gao et al., 2022). The idea that Pou5f3 and Nanog is competing at a same binding site and may function as a blocker is indeed interesting and advancing our knowledge, but molecular experimental evidence is lacking. How Pou5f3 and Nanog function as a non- pioneer blocker is also completely a blackbox. To address underlying molecular mechanisms, at least the authors should do experiments such as gel- shift assay to enhancer sequences of specific group of genes.  
+
+The authors claim the presence of additional genome activators, but the evidence is not enough. First, in Fig. 2a, why did not the authors use the MZsox19b for the classification? Second, the results show that \(\sim 50\%\) or \(\sim 80\%\) of group 4. - TdARs can be rescued by Sox19b alone, or by Sox19b and Nanog, respectively (Fig S3). From these data, it is likely that SoxB, or SoxB and Nanog is required on chromatin accessibility of most of group 4. - TdARs.  
+
+Characterization of antagonistic enhancers is not enough. They show GC content for 1. PN, 2. P, 3. N, and 4. -, but comparison between p- n+, p+ n- and others is required. Furthermore, they mention that exact match to the motif is important for Pou5f3 and Nanog to function as pioneer TFs, but this does not explain the difference between synergistic and antagonistic enhancers. Why do Pou5f3 and Nanog function as blocker only in antagonistic enhancers, but not in synergistic enhancers? The author also need to have additional experimental data as stated above.  
+
+Regarding the mathematical modelling; in Figure 5c, expressions of 3b. P- N+ genes do not seem to be upregulated in MZps. They even seems down regulated at stages before 6 hpf. Furthermore, gene ontology analyses are not enough to evaluate the model accuracy, and validation in other method is required. For example, is there any differences in sequence feature between enhancers of the six groups sorted by the model?  
+
+Reviewer #2 (Remarks to the Author):  
+
+The manuscript "Activator- blocker model of transcriptional regulation by pioneer- like factors," by Riesle et al takes a deep dive into functions for the transcription factors Pou5f3, Sox19b and Nanog during zebrafish zygotic genome activation. Their analysis uses a combination of elegant genetic and genomic approaches. The results reveal surprising complexities in the ways in which Pou5f3, Sox19b and Nanog work both together and antagonistically to regulate early transcription.
+
+<--- Page Split --->
+
+Overall this is a rigorous study which produces useful data sets and shifts the way we need to think about how these transcription factors act during zebrafish ZGA. I have only minor comments.  
+
+The authors do a very nice job of detailing the crosses performed to generate the various mutant lines in the methods, and the original alleles are cited. However, given the importance of the mutants it might be worth reviewing the nature of the alleles in the start of the results section. For example, do any of them produce proteins that could still have some DNA binding capacity?  
+
+Figure 2b, not clear exactly what control genomic regions are.  
+
+With respect to figure 2, it could be more helpful to include a little more detail on the binning of up down and unchanged groups in the text or figure legend (what were the cutoffs/criteria for including). I similarly struggled to find this information in the methods, if it is there it is not easy to find.  
+
+Figure 3 panel 1 is a little confusing. Is purple color code also needed?  
+
+With respect to figure 3 F, it isn't entirely clear to me why this group couldn't reflect a requirement for all three TFs together, rather than any one plus GC binding factor- Fig 3b seems to suggests binding motifs for all three are detected in this group, although the so: pou motif is underrepresented. The GC enrichment could still represent a hypothetical GC binding factor required in addition to the three factors? Alternatively, is it possible that there is simply a structural property of these GC rich regions that is keeping them more open? I believe there is reasonable data to support a relationship between GC content and nucleosome occupancy  
+
+The rescue experiments in Fig S3 are quite nice. I almost wonder if they might be better in the main text of the manuscript.  
+
+Figure 4, Panel C in particular is really important and very difficult to interpret, not at all intuitive. Is there any other way to present this data more clearly? I had to sit with the browser shots in panel H a good while to understand what was going on holistically in figure 4.  
+
+I think a prediction of the model in 4f is that you should never be able to IP A and B from chromatin together at these types of sites. Is it possible to do sequential IPs to explore that possibility? This could be interesting additional support, but if this is technically too challenging, could be viewed as beyond the current paper's scope.
+
+<--- Page Split --->
+
+Dear reviewers #1 and #2,  
+
+We are grateful that you professionally spotted the weak points in our manuscript, which helped us to improve it. We performed additional experiments, added new sub- chapter in the results section with new main Fig. 5 and Fig. S6, performed additional analysis and made changes in most of the figures. We numbered your comments and linked them as comments to the yellow- marked text in the PDF file "Riesle_main_and SUPPL_marked for reviewers", which contains main text and figures, supplementary figures and legends and also additional supplementary Figure for Reviewer 1, at the end of the supplementary material. Please find detailed answers to your criticisms below. Sincerely, the authors.  
+
+## REVIEWER COMMENTS  
+
+Reviewer #1 (Remarks to the Author):  
+
+In this manuscript, Riesle and Gao et al. analyzed ATAC- seq, H3K27ac ChIP- seq, and transcriptome of zebrafish mutants of ZGA regulator TFs. They found the existence of synergistic and antagonistic enhancer types, and Pou5f3 and Nanog can function as activator or blocker in antagonistic enhancers. The topic is unquestionably important for the field, but this reviewer feels that the novelty of the study seems weak. I have several major concerns as described below.  
+
+R1- 1. Regarding the novelty, the fact that the pioneer factors (Pou5f3 and Sox19b) is preventing premature expression for some of the genes during ZGA has already been reported by the same group (Gao et al., 2022). The idea that Pou5f3 and Nanog is competing at a same binding site and may function as a blocker is indeed interesting and advancing our knowledge, but molecular experimental evidence is lacking. How Pou5f3 and Nanog function as a non- pioneer blocker is also completely a blackbox. To address underlying molecular mechanisms, at least the authors should do experiments such as gel- shift assay to enhancer sequences of specific group of genes.  
+
+A 1- 1. We agree with the comment and provide requested evidence. We performed series of the gel- shift assays with the oligos from different enhancers, added the new figures 5 and S6 and a chapter "Pou5f3 and Nanog bind to the common binding sites in a mutually exclusive way". We also show that in most cases TdARs have only one match to either Pou5f3 or Nanog motif (new panel Fig. S5f).  
+
+R1- 2. The authors claim the presence of additional genome activators, but the evidence is not enough. First, in Fig. 2a, why did not the authors use the MZsox19b for the classification? Second, the results show that \(\sim 50\%\) or \(\sim 80\%\) of group 4- TdARs can be rescued by Sox19b alone, or by Sox19b and Nanog, respectively (Fig S3). From these data, it is likely that SoxB, or SoxB and Nanog is required on chromatin accessibility of most of group 4- TdARs.  
+
+A1- 2. We agree with both points. In response to this criticism we - made additional analysis of 4- group, using ATAC- seq in MZsox19b and three double mutants (new panel c in the Fig. S3). - removed "hypothetical GC protein" from the scheme in Fig.3; and speculations about "hypothetical GC protein" from the corresponding text in the results.  
+
+R1- 3. Characterization of antagonistic enhancers is not enough. They show GC content for 1. PN, 2. P, 3. N, and 4. -, but comparison between p- n+, p+ n- and others is required.  
+
+A1- 3. Done (Fig. S4). Conclusion: overall GC content of enhancers activated by Pou5f3 is lower than of the enhancers activated by Nanog. This is seen in both ATAC- seq (Fig.3e, Fig. S4b) and H3K27ac comparisons between TdARs (Fig. S4b).  
+
+R1- 4. Furthermore, they mention that exact match to the motif is important for Pou5f3 and Nanog to function as pioneer TFs, but this does not explain the difference between synergistic and antagonistic enhancers. Why do Pou5f3 and Nanog function as blocker only in antagonistic enhancers, but not in synergistic enhancers? The author also need to have additional experimental data as stated above.  
+
+A1- 4. There are two parts of the answer to this criticism:  
+
+1) Yes, we provided additional experimental data and confirmed now with gel-shifts, that the factor which binds stronger in-vitro works as an activator in-vivo (9 out of 9 oligos with single motif, where the binding worked for any of the two TFs, see the new Fig. 5 panel e). We also show that GC content is important: Pou5f3 activates the enhancers with lower GC content range, than Nanog, as judged by pioneer activity (Fig.3e) and H3K27ac change (Fig.S4).
+
+<--- Page Split --->
+
+2) No, we can not explain the mechanistic difference between synergistic and antagonistic enhancers from gel-shifts and bulk genomic data. We assume that cell-specific cofactors are involved (see discussion). Single-cell analysis is required to answer this question; as far as we know ChIP-seq technique for single cells is not yet developed. We hope that our manuscript is novel enough to be published without it.  
+
+R1- 5. Regarding the mathematical modelling; in Figure 5c, expressions of 3b. P- N+ genes do not seem to be upregulated in MZps. They even seems down regulated at stages before 6 hpf.f  
+
+This is because most of 3b. P- N+ genes are coactivated by SOXB1 sum, and SOXB1 sum is decreased in MZps. ( see the supplementary Figure for Reviewer 1 included in the file for reviewers). 3b. P- N+ model group consists of three mini- models: of three mini- models: S+P- N+ (best fit to 398 transcripts), SOP- N+ (best fit to 24 transcripts) and S- P- N+ (best fit to 6 transcripts). In response to this criticism we changed the summatory figure S8, which shows now the example fits to all mini- models ( and not to the groups as before).  
+
+R1- 6. Furthermore, gene ontology analyses are not enough to evaluate the model accuracy, and validation in other method is required.  
+
+We did not intent to validate the modeling with GO analysis. In response to this criticism we completly removed this GO analysis from the paper, not to distract the reader attention. We renamed the sub- chapters and put the cross- validation part (the synergistically and antagonistically regulated transcripts are linked to synergistically and antagonistically regulated enhancers) just after the description of the modeling results (see the marking of the text for reviewers).  
+
+R1- 7. For example, is there any differences in sequence feature between enhancers of the six groups sorted by the model?  
+
+Yes, there is a difference in both sequence features: motif frequency and in GC content. We show it now in Fig. S9 c,d, as additional validation.  
+
+Reviewer #2 (Remarks to the Author):  
+
+The manuscript "Activator- blocker model of transcriptional regulation by pioneer- like factors," by Riesle et al takes a deep dive into functions for the transcription factors Pou53, Sox19b and Nanog during zebrafish zygotic genome activation. Their analysis uses a combination of elegant genetic and genomic approaches. The results reveal surprising complexities in the ways in which Pou53, Sox19b and Nanog work both together and antagonistically to regulate early transcription.  
+
+Overall this is a rigorous study which produces useful data sets and shifts the way we need to think about how these transcription factors act during zebrafish ZGA. I have only minor comments.  
+
+R2- 1. The authors do a very nice job of detailing the crosses performed to generate the various mutant lines in the methods, and the original alleles are cited. However, given the importance of the mutants it might be worth reviewing the nature of the alleles in the start of the results section. For example, do any of them produce proteins that could still have some DNA binding capacity?  
+
+A2- 1. All the mutants are genetic nulls: Pou53 mutant has a point- mutation before DNA- binding domain, Sox19b and Nanog have TALEN- induced frameshifts before the DNA- binding domains. We mention it in the results and give the references to the original publications where it was characterized after each mutant name.  
+
+R2- 2. Figure 2b, not clear exactly what control genomic regions are.  
+
+A2- 2. Included in the legend for Figure 2b: "To obtain control genomic regions (co, dotted line), genomic coordinates of all ARs were shifted 1 kb downstream".  
+
+R2- 3. With respect to figure 2, it could be more helpful to include a little more detail on the binning of up down and unchanged groups in the text or figure legend (what were the cutoffs/criteria for including). I similarly struggled to find this information in the methods, if it is there it is not easy to find.
+
+<--- Page Split --->
+
+A2- 3. We agree (it is only in "Methods" and in too detailed). We included in the figure legend for Fig. 2a: "Three groups of accessible regions (ARs) were selected as follows: in "down" and "up" regions, ATAC- signal was reduced or increased, respectively, in six MZtriple biological replicates compared to seven wild- type biological replicates with false discovery rate \(< 5\%\) . "same" - the remaining ARs which were considered unchanged". R2- 4. Figure 3 panel 1 is a little confusing. Is purple color code also needed?  
+
+A2- 4. Yes! We included it to Fig 3a and to Fig S3a.  
+
+R2- 5. With respect to figure 3 F, it isn't entirely clear to me why this group couldn't reflect a requirement for all three TFs together, rather than any one plus GC binding factor- Fig 3b seems to suggests binding motifs for all three are detected in this group, although the sox: pou motif is underrepresented. The GC enrichment could still represent a hypothetical GC binding factor required in addition to the three factors? Alternatively, is it possible that there is simply a structural property of these GC rich regions that is keeping them more open? I believe there is reasonable data to support a relationship between GC content and nucleosome occupancy  
+
+A2- 5. We agree, we do not really need the GC factor in the scheme. We replaced it, removed the speculations about GC protein from the text, and included additional analysis of the 4. - group to the Figure S3c.  
+
+R2- 6. The rescue experiments in Fig S3 are quite nice. I almost wonder if they might be better in the main text of the manuscript.  
+
+A2- 6. We agree. We moved the experiments from the Fig.S3 to Fig. 3, panels f,g.  
+
+R2- 7. Figure 4, Panel C in particular is really important and very difficult to interpret, not at all intuitive. Is there any other way to present this data more clearly? I had to sit with the browser shots in panel H a good while to understand what was going on holistically in figure 4.  
+
+A2- 7. We tried our best to make the things simpler to precept. As a result, we expanded the analysis, the panel C is now moved from Fig. 4 to the new Supplementary Fig.4 which is completely filled with analysis. The order of remaining panels in the main Fig. 4 is reorganized.  
+
+R2- 8. I think a prediction of the model in 4f is that you should never be able to IP A and B from chromatin together at these types of sites. Is it possible to do sequential IPs to explore that possibility? This could be interesting additional support, but if this is technically too challenging, could be viewed as beyond the current paper's scope.  
+
+A2- 8. Thank you for 1) the clear formulation of this very important message, and 2) for putting this request in the minor comments. We show now in the new panel of the supplementary Fig. 5f that most of the open regions bound by TFs have only one site, either for Pou5f3 or for Nanog. We have also done a set of gel retardation assays with tagged Pou5f3 and Nanog and supershifts, showing that Pou5f3 and Nanog do not bind together to the same DNA motifs in- vitro. We included the new Fig 5, Fig. S5 and new sub- chapter of the results: "Pou5f3 and Nanog bind to the common binding sites in a mutually exclusive way".
+
+<--- Page Split --->
+
+## REVIEWER COMMENTS  
+
+Reviewer #1 (Remarks to the Author):  
+
+Reviewer #1 (Remarks to the Author):The reviewer appreciates the authors' efforts in incorporating new data into the revised manuscript. However, I still have a concern. The authors have made an attempt to demonstrate the mutually exclusive binding of Pou5f3 and Nanog using gel- shift assays (Figure 5). Nevertheless, the data only indicate that either Pou5f3 or Nanog strongly binds, while the other protein binds weakly to the same oligo. It does not clearly show the blocking of one protein by the other. It would be valuable for the authors to provide evidence by demonstrating the actual replacement of Pou5f3 binding with Nanog binding as the concentration of Nanog increases and vice versa, using the gel- shift assay the authors established. The reviewer is sure that the authors would easily perform this kind of experiments. Given that the activator- blocker model is one of the most crucial assertions of this paper, the reviewer thinks, it is essential to experimentally validate the molecular mechanism.  
+
+Reviewer #2 (Remarks to the Author):  
+
+All of my concerns regarding this manuscript have been addressed.
+
+<--- Page Split --->
+
+## REVIEWER COMMENTS  
+
+Reviewer #1 (Remarks to the Author):  
+
+R1- 1  
+
+The reviewer appreciates the authors' efforts in incorporating new data into the revised manuscript. However, I still have a concern. The authors have made an attempt to demonstrate the mutually exclusive binding of Pou5f3 and Nanog using gel- shift assays (Figure 5). Nevertheless, the data only indicate that either Pou5f3 or Nanog strongly binds, while the other protein binds weakly to the same oligo. It does not clearly show the blocking of one protein by the other. It would be valuable for the authors to provide evidence by demonstrating the actual replacement of Pou5f3 binding with Nanog binding as the concentration of Nanog increases and vice versa, using the gel- shift assay the authors established. The reviewer is sure that the authors would easily perform this kind of experiments. Given that the activator- blocker model is one of the most crucial assertions of this paper, the reviewer thinks, it is essential to experimentally validate the molecular mechanism.  
+
+I performed the requested EMSA experiments, included them as a new Supplementary Fig.7 and referred them in the text (please see the yellow marking in the PDF file for reviewers).  
+
+Sincerely, Dr.Daria Onichtchouk.  
+
+Reviewer #2 (Remarks to the Author):  
+
+All of my concerns regarding this manuscript have been addressed.
+
+<--- Page Split --->

peer_reviews/129058fcec31e99c0819f8921d0245dbb3bd7403d48f93401d336635a4c8ebb2/supplementary_0_Peer Review File/supplementary_0_Peer Review File_det.mmd

@@ -0,0 +1,232 @@
+<|ref|>title<|/ref|><|det|>[[100, 40, 506, 90]]<|/det|>
+# nature portfolio  
+
+<|ref|>text<|/ref|><|det|>[[106, 110, 373, 139]]<|/det|>
+Peer Review File  
+
+<|ref|>text<|/ref|><|det|>[[108, 162, 808, 219]]<|/det|>
+Activator- blocker model of transcriptional regulation by pioneer- like factors  
+
+<|ref|>image<|/ref|><|det|>[[95, 732, 262, 780]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[271, 732, 880, 784]]<|/det|>
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[116, 88, 305, 104]]<|/det|>
+## REVIEWER COMMENTS  
+
+<|ref|>text<|/ref|><|det|>[[116, 118, 404, 134]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 148, 880, 224]]<|/det|>
+In this manuscript, Riesle and Gao et al. analyzed ATAC- seq, H3K27ac ChIP- seq, and transcriptome of zebrafish mutants of ZGA regulator TFs. They found the existence of synergistic and antagonistic enhancer types, and Pou5f3 and Nanog can function as activator or blocker in antagonistic enhancers. The topic is unquestionably important for the field, but this reviewer feels that the novelty of the study seems weak. I have several major concerns as described below.  
+
+<|ref|>text<|/ref|><|det|>[[115, 252, 879, 358]]<|/det|>
+Regarding the novelty, the fact that the pioneer factors (Pou5f3 and Sox19b) is preventing premature expression for some of the genes during ZGA has already been reported by the same group (Gao et al., 2022). The idea that Pou5f3 and Nanog is competing at a same binding site and may function as a blocker is indeed interesting and advancing our knowledge, but molecular experimental evidence is lacking. How Pou5f3 and Nanog function as a non- pioneer blocker is also completely a blackbox. To address underlying molecular mechanisms, at least the authors should do experiments such as gel- shift assay to enhancer sequences of specific group of genes.  
+
+<|ref|>text<|/ref|><|det|>[[115, 386, 877, 462]]<|/det|>
+The authors claim the presence of additional genome activators, but the evidence is not enough. First, in Fig. 2a, why did not the authors use the MZsox19b for the classification? Second, the results show that \(\sim 50\%\) or \(\sim 80\%\) of group 4. - TdARs can be rescued by Sox19b alone, or by Sox19b and Nanog, respectively (Fig S3). From these data, it is likely that SoxB, or SoxB and Nanog is required on chromatin accessibility of most of group 4. - TdARs.  
+
+<|ref|>text<|/ref|><|det|>[[115, 490, 875, 581]]<|/det|>
+Characterization of antagonistic enhancers is not enough. They show GC content for 1. PN, 2. P, 3. N, and 4. -, but comparison between p- n+, p+ n- and others is required. Furthermore, they mention that exact match to the motif is important for Pou5f3 and Nanog to function as pioneer TFs, but this does not explain the difference between synergistic and antagonistic enhancers. Why do Pou5f3 and Nanog function as blocker only in antagonistic enhancers, but not in synergistic enhancers? The author also need to have additional experimental data as stated above.  
+
+<|ref|>text<|/ref|><|det|>[[115, 609, 880, 685]]<|/det|>
+Regarding the mathematical modelling; in Figure 5c, expressions of 3b. P- N+ genes do not seem to be upregulated in MZps. They even seems down regulated at stages before 6 hpf. Furthermore, gene ontology analyses are not enough to evaluate the model accuracy, and validation in other method is required. For example, is there any differences in sequence feature between enhancers of the six groups sorted by the model?  
+
+<|ref|>text<|/ref|><|det|>[[116, 744, 404, 759]]<|/det|>
+Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 789, 867, 864]]<|/det|>
+The manuscript "Activator- blocker model of transcriptional regulation by pioneer- like factors," by Riesle et al takes a deep dive into functions for the transcription factors Pou5f3, Sox19b and Nanog during zebrafish zygotic genome activation. Their analysis uses a combination of elegant genetic and genomic approaches. The results reveal surprising complexities in the ways in which Pou5f3, Sox19b and Nanog work both together and antagonistically to regulate early transcription.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 89, 857, 119]]<|/det|>
+Overall this is a rigorous study which produces useful data sets and shifts the way we need to think about how these transcription factors act during zebrafish ZGA. I have only minor comments.  
+
+<|ref|>text<|/ref|><|det|>[[115, 148, 878, 209]]<|/det|>
+The authors do a very nice job of detailing the crosses performed to generate the various mutant lines in the methods, and the original alleles are cited. However, given the importance of the mutants it might be worth reviewing the nature of the alleles in the start of the results section. For example, do any of them produce proteins that could still have some DNA binding capacity?  
+
+<|ref|>text<|/ref|><|det|>[[115, 223, 576, 238]]<|/det|>
+Figure 2b, not clear exactly what control genomic regions are.  
+
+<|ref|>text<|/ref|><|det|>[[115, 253, 875, 298]]<|/det|>
+With respect to figure 2, it could be more helpful to include a little more detail on the binning of up down and unchanged groups in the text or figure legend (what were the cutoffs/criteria for including). I similarly struggled to find this information in the methods, if it is there it is not easy to find.  
+
+<|ref|>text<|/ref|><|det|>[[115, 312, 640, 327]]<|/det|>
+Figure 3 panel 1 is a little confusing. Is purple color code also needed?  
+
+<|ref|>text<|/ref|><|det|>[[115, 342, 881, 448]]<|/det|>
+With respect to figure 3 F, it isn't entirely clear to me why this group couldn't reflect a requirement for all three TFs together, rather than any one plus GC binding factor- Fig 3b seems to suggests binding motifs for all three are detected in this group, although the so: pou motif is underrepresented. The GC enrichment could still represent a hypothetical GC binding factor required in addition to the three factors? Alternatively, is it possible that there is simply a structural property of these GC rich regions that is keeping them more open? I believe there is reasonable data to support a relationship between GC content and nucleosome occupancy  
+
+<|ref|>text<|/ref|><|det|>[[115, 462, 866, 492]]<|/det|>
+The rescue experiments in Fig S3 are quite nice. I almost wonder if they might be better in the main text of the manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[115, 506, 877, 551]]<|/det|>
+Figure 4, Panel C in particular is really important and very difficult to interpret, not at all intuitive. Is there any other way to present this data more clearly? I had to sit with the browser shots in panel H a good while to understand what was going on holistically in figure 4.  
+
+<|ref|>text<|/ref|><|det|>[[115, 566, 861, 625]]<|/det|>
+I think a prediction of the model in 4f is that you should never be able to IP A and B from chromatin together at these types of sites. Is it possible to do sequential IPs to explore that possibility? This could be interesting additional support, but if this is technically too challenging, could be viewed as beyond the current paper's scope.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[117, 81, 295, 95]]<|/det|>
+Dear reviewers #1 and #2,  
+
+<|ref|>text<|/ref|><|det|>[[116, 105, 866, 205]]<|/det|>
+We are grateful that you professionally spotted the weak points in our manuscript, which helped us to improve it. We performed additional experiments, added new sub- chapter in the results section with new main Fig. 5 and Fig. S6, performed additional analysis and made changes in most of the figures. We numbered your comments and linked them as comments to the yellow- marked text in the PDF file "Riesle_main_and SUPPL_marked for reviewers", which contains main text and figures, supplementary figures and legends and also additional supplementary Figure for Reviewer 1, at the end of the supplementary material. Please find detailed answers to your criticisms below. Sincerely, the authors.  
+
+<|ref|>sub_title<|/ref|><|det|>[[117, 216, 291, 230]]<|/det|>
+## REVIEWER COMMENTS  
+
+<|ref|>text<|/ref|><|det|>[[118, 241, 372, 255]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[117, 265, 876, 329]]<|/det|>
+In this manuscript, Riesle and Gao et al. analyzed ATAC- seq, H3K27ac ChIP- seq, and transcriptome of zebrafish mutants of ZGA regulator TFs. They found the existence of synergistic and antagonistic enhancer types, and Pou5f3 and Nanog can function as activator or blocker in antagonistic enhancers. The topic is unquestionably important for the field, but this reviewer feels that the novelty of the study seems weak. I have several major concerns as described below.  
+
+<|ref|>text<|/ref|><|det|>[[117, 351, 860, 439]]<|/det|>
+R1- 1. Regarding the novelty, the fact that the pioneer factors (Pou5f3 and Sox19b) is preventing premature expression for some of the genes during ZGA has already been reported by the same group (Gao et al., 2022). The idea that Pou5f3 and Nanog is competing at a same binding site and may function as a blocker is indeed interesting and advancing our knowledge, but molecular experimental evidence is lacking. How Pou5f3 and Nanog function as a non- pioneer blocker is also completely a blackbox. To address underlying molecular mechanisms, at least the authors should do experiments such as gel- shift assay to enhancer sequences of specific group of genes.  
+
+<|ref|>text<|/ref|><|det|>[[117, 449, 880, 502]]<|/det|>
+A 1- 1. We agree with the comment and provide requested evidence. We performed series of the gel- shift assays with the oligos from different enhancers, added the new figures 5 and S6 and a chapter "Pou5f3 and Nanog bind to the common binding sites in a mutually exclusive way". We also show that in most cases TdARs have only one match to either Pou5f3 or Nanog motif (new panel Fig. S5f).  
+
+<|ref|>text<|/ref|><|det|>[[117, 526, 880, 591]]<|/det|>
+R1- 2. The authors claim the presence of additional genome activators, but the evidence is not enough. First, in Fig. 2a, why did not the authors use the MZsox19b for the classification? Second, the results show that \(\sim 50\%\) or \(\sim 80\%\) of group 4- TdARs can be rescued by Sox19b alone, or by Sox19b and Nanog, respectively (Fig S3). From these data, it is likely that SoxB, or SoxB and Nanog is required on chromatin accessibility of most of group 4- TdARs.  
+
+<|ref|>text<|/ref|><|det|>[[116, 601, 870, 666]]<|/det|>
+A1- 2. We agree with both points. In response to this criticism we - made additional analysis of 4- group, using ATAC- seq in MZsox19b and three double mutants (new panel c in the Fig. S3). - removed "hypothetical GC protein" from the scheme in Fig.3; and speculations about "hypothetical GC protein" from the corresponding text in the results.  
+
+<|ref|>text<|/ref|><|det|>[[116, 686, 863, 714]]<|/det|>
+R1- 3. Characterization of antagonistic enhancers is not enough. They show GC content for 1. PN, 2. P, 3. N, and 4. -, but comparison between p- n+, p+ n- and others is required.  
+
+<|ref|>text<|/ref|><|det|>[[116, 724, 848, 763]]<|/det|>
+A1- 3. Done (Fig. S4). Conclusion: overall GC content of enhancers activated by Pou5f3 is lower than of the enhancers activated by Nanog. This is seen in both ATAC- seq (Fig.3e, Fig. S4b) and H3K27ac comparisons between TdARs (Fig. S4b).  
+
+<|ref|>text<|/ref|><|det|>[[116, 773, 863, 825]]<|/det|>
+R1- 4. Furthermore, they mention that exact match to the motif is important for Pou5f3 and Nanog to function as pioneer TFs, but this does not explain the difference between synergistic and antagonistic enhancers. Why do Pou5f3 and Nanog function as blocker only in antagonistic enhancers, but not in synergistic enhancers? The author also need to have additional experimental data as stated above.  
+
+<|ref|>text<|/ref|><|det|>[[116, 835, 491, 849]]<|/det|>
+A1- 4. There are two parts of the answer to this criticism:  
+
+<|ref|>text<|/ref|><|det|>[[145, 849, 874, 912]]<|/det|>
+1) Yes, we provided additional experimental data and confirmed now with gel-shifts, that the factor which binds stronger in-vitro works as an activator in-vivo (9 out of 9 oligos with single motif, where the binding worked for any of the two TFs, see the new Fig. 5 panel e). We also show that GC content is important: Pou5f3 activates the enhancers with lower GC content range, than Nanog, as judged by pioneer activity (Fig.3e) and H3K27ac change (Fig.S4).
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[145, 81, 877, 134]]<|/det|>
+2) No, we can not explain the mechanistic difference between synergistic and antagonistic enhancers from gel-shifts and bulk genomic data. We assume that cell-specific cofactors are involved (see discussion). Single-cell analysis is required to answer this question; as far as we know ChIP-seq technique for single cells is not yet developed. We hope that our manuscript is novel enough to be published without it.  
+
+<|ref|>text<|/ref|><|det|>[[118, 155, 850, 182]]<|/det|>
+R1- 5. Regarding the mathematical modelling; in Figure 5c, expressions of 3b. P- N+ genes do not seem to be upregulated in MZps. They even seems down regulated at stages before 6 hpf.f  
+
+<|ref|>text<|/ref|><|det|>[[117, 192, 880, 260]]<|/det|>
+This is because most of 3b. P- N+ genes are coactivated by SOXB1 sum, and SOXB1 sum is decreased in MZps. ( see the supplementary Figure for Reviewer 1 included in the file for reviewers). 3b. P- N+ model group consists of three mini- models: of three mini- models: S+P- N+ (best fit to 398 transcripts), SOP- N+ (best fit to 24 transcripts) and S- P- N+ (best fit to 6 transcripts). In response to this criticism we changed the summatory figure S8, which shows now the example fits to all mini- models ( and not to the groups as before).  
+
+<|ref|>text<|/ref|><|det|>[[118, 290, 850, 316]]<|/det|>
+R1- 6. Furthermore, gene ontology analyses are not enough to evaluate the model accuracy, and validation in other method is required.  
+
+<|ref|>text<|/ref|><|det|>[[117, 326, 880, 391]]<|/det|>
+We did not intent to validate the modeling with GO analysis. In response to this criticism we completly removed this GO analysis from the paper, not to distract the reader attention. We renamed the sub- chapters and put the cross- validation part (the synergistically and antagonistically regulated transcripts are linked to synergistically and antagonistically regulated enhancers) just after the description of the modeling results (see the marking of the text for reviewers).  
+
+<|ref|>text<|/ref|><|det|>[[117, 412, 861, 439]]<|/det|>
+R1- 7. For example, is there any differences in sequence feature between enhancers of the six groups sorted by the model?  
+
+<|ref|>text<|/ref|><|det|>[[116, 450, 863, 477]]<|/det|>
+Yes, there is a difference in both sequence features: motif frequency and in GC content. We show it now in Fig. S9 c,d, as additional validation.  
+
+<|ref|>text<|/ref|><|det|>[[118, 535, 372, 549]]<|/det|>
+Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[117, 572, 877, 636]]<|/det|>
+The manuscript "Activator- blocker model of transcriptional regulation by pioneer- like factors," by Riesle et al takes a deep dive into functions for the transcription factors Pou53, Sox19b and Nanog during zebrafish zygotic genome activation. Their analysis uses a combination of elegant genetic and genomic approaches. The results reveal surprising complexities in the ways in which Pou53, Sox19b and Nanog work both together and antagonistically to regulate early transcription.  
+
+<|ref|>text<|/ref|><|det|>[[118, 658, 857, 685]]<|/det|>
+Overall this is a rigorous study which produces useful data sets and shifts the way we need to think about how these transcription factors act during zebrafish ZGA. I have only minor comments.  
+
+<|ref|>text<|/ref|><|det|>[[117, 707, 880, 759]]<|/det|>
+R2- 1. The authors do a very nice job of detailing the crosses performed to generate the various mutant lines in the methods, and the original alleles are cited. However, given the importance of the mutants it might be worth reviewing the nature of the alleles in the start of the results section. For example, do any of them produce proteins that could still have some DNA binding capacity?  
+
+<|ref|>text<|/ref|><|det|>[[117, 770, 873, 809]]<|/det|>
+A2- 1. All the mutants are genetic nulls: Pou53 mutant has a point- mutation before DNA- binding domain, Sox19b and Nanog have TALEN- induced frameshifts before the DNA- binding domains. We mention it in the results and give the references to the original publications where it was characterized after each mutant name.  
+
+<|ref|>text<|/ref|><|det|>[[118, 819, 572, 832]]<|/det|>
+R2- 2. Figure 2b, not clear exactly what control genomic regions are.  
+
+<|ref|>text<|/ref|><|det|>[[117, 843, 812, 869]]<|/det|>
+A2- 2. Included in the legend for Figure 2b: "To obtain control genomic regions (co, dotted line), genomic coordinates of all ARs were shifted 1 kb downstream".  
+
+<|ref|>text<|/ref|><|det|>[[117, 880, 880, 918]]<|/det|>
+R2- 3. With respect to figure 2, it could be more helpful to include a little more detail on the binning of up down and unchanged groups in the text or figure legend (what were the cutoffs/criteria for including). I similarly struggled to find this information in the methods, if it is there it is not easy to find.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[116, 81, 866, 147]]<|/det|>
+A2- 3. We agree (it is only in "Methods" and in too detailed). We included in the figure legend for Fig. 2a: "Three groups of accessible regions (ARs) were selected as follows: in "down" and "up" regions, ATAC- signal was reduced or increased, respectively, in six MZtriple biological replicates compared to seven wild- type biological replicates with false discovery rate \(< 5\%\) . "same" - the remaining ARs which were considered unchanged". R2- 4. Figure 3 panel 1 is a little confusing. Is purple color code also needed?  
+
+<|ref|>text<|/ref|><|det|>[[118, 147, 580, 160]]<|/det|>
+A2- 4. Yes! We included it to Fig 3a and to Fig S3a.  
+
+<|ref|>text<|/ref|><|det|>[[116, 168, 883, 244]]<|/det|>
+R2- 5. With respect to figure 3 F, it isn't entirely clear to me why this group couldn't reflect a requirement for all three TFs together, rather than any one plus GC binding factor- Fig 3b seems to suggests binding motifs for all three are detected in this group, although the sox: pou motif is underrepresented. The GC enrichment could still represent a hypothetical GC binding factor required in addition to the three factors? Alternatively, is it possible that there is simply a structural property of these GC rich regions that is keeping them more open? I believe there is reasonable data to support a relationship between GC content and nucleosome occupancy  
+
+<|ref|>text<|/ref|><|det|>[[116, 253, 860, 280]]<|/det|>
+A2- 5. We agree, we do not really need the GC factor in the scheme. We replaced it, removed the speculations about GC protein from the text, and included additional analysis of the 4. - group to the Figure S3c.  
+
+<|ref|>text<|/ref|><|det|>[[116, 290, 868, 317]]<|/det|>
+R2- 6. The rescue experiments in Fig S3 are quite nice. I almost wonder if they might be better in the main text of the manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[116, 327, 661, 341]]<|/det|>
+A2- 6. We agree. We moved the experiments from the Fig.S3 to Fig. 3, panels f,g.  
+
+<|ref|>text<|/ref|><|det|>[[116, 352, 857, 390]]<|/det|>
+R2- 7. Figure 4, Panel C in particular is really important and very difficult to interpret, not at all intuitive. Is there any other way to present this data more clearly? I had to sit with the browser shots in panel H a good while to understand what was going on holistically in figure 4.  
+
+<|ref|>text<|/ref|><|det|>[[116, 401, 875, 440]]<|/det|>
+A2- 7. We tried our best to make the things simpler to precept. As a result, we expanded the analysis, the panel C is now moved from Fig. 4 to the new Supplementary Fig.4 which is completely filled with analysis. The order of remaining panels in the main Fig. 4 is reorganized.  
+
+<|ref|>text<|/ref|><|det|>[[116, 450, 878, 488]]<|/det|>
+R2- 8. I think a prediction of the model in 4f is that you should never be able to IP A and B from chromatin together at these types of sites. Is it possible to do sequential IPs to explore that possibility? This could be interesting additional support, but if this is technically too challenging, could be viewed as beyond the current paper's scope.  
+
+<|ref|>text<|/ref|><|det|>[[116, 498, 870, 574]]<|/det|>
+A2- 8. Thank you for 1) the clear formulation of this very important message, and 2) for putting this request in the minor comments. We show now in the new panel of the supplementary Fig. 5f that most of the open regions bound by TFs have only one site, either for Pou5f3 or for Nanog. We have also done a set of gel retardation assays with tagged Pou5f3 and Nanog and supershifts, showing that Pou5f3 and Nanog do not bind together to the same DNA motifs in- vitro. We included the new Fig 5, Fig. S5 and new sub- chapter of the results: "Pou5f3 and Nanog bind to the common binding sites in a mutually exclusive way".
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[116, 89, 305, 104]]<|/det|>
+## REVIEWER COMMENTS  
+
+<|ref|>text<|/ref|><|det|>[[116, 119, 404, 134]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 148, 877, 298]]<|/det|>
+Reviewer #1 (Remarks to the Author):The reviewer appreciates the authors' efforts in incorporating new data into the revised manuscript. However, I still have a concern. The authors have made an attempt to demonstrate the mutually exclusive binding of Pou5f3 and Nanog using gel- shift assays (Figure 5). Nevertheless, the data only indicate that either Pou5f3 or Nanog strongly binds, while the other protein binds weakly to the same oligo. It does not clearly show the blocking of one protein by the other. It would be valuable for the authors to provide evidence by demonstrating the actual replacement of Pou5f3 binding with Nanog binding as the concentration of Nanog increases and vice versa, using the gel- shift assay the authors established. The reviewer is sure that the authors would easily perform this kind of experiments. Given that the activator- blocker model is one of the most crucial assertions of this paper, the reviewer thinks, it is essential to experimentally validate the molecular mechanism.  
+
+<|ref|>text<|/ref|><|det|>[[116, 327, 404, 342]]<|/det|>
+Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 357, 616, 372]]<|/det|>
+All of my concerns regarding this manuscript have been addressed.
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[118, 105, 291, 119]]<|/det|>
+## REVIEWER COMMENTS  
+
+<|ref|>text<|/ref|><|det|>[[118, 131, 373, 145]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[118, 156, 152, 168]]<|/det|>
+R1- 1  
+
+<|ref|>text<|/ref|><|det|>[[117, 168, 878, 280]]<|/det|>
+The reviewer appreciates the authors' efforts in incorporating new data into the revised manuscript. However, I still have a concern. The authors have made an attempt to demonstrate the mutually exclusive binding of Pou5f3 and Nanog using gel- shift assays (Figure 5). Nevertheless, the data only indicate that either Pou5f3 or Nanog strongly binds, while the other protein binds weakly to the same oligo. It does not clearly show the blocking of one protein by the other. It would be valuable for the authors to provide evidence by demonstrating the actual replacement of Pou5f3 binding with Nanog binding as the concentration of Nanog increases and vice versa, using the gel- shift assay the authors established. The reviewer is sure that the authors would easily perform this kind of experiments. Given that the activator- blocker model is one of the most crucial assertions of this paper, the reviewer thinks, it is essential to experimentally validate the molecular mechanism.  
+
+<|ref|>text<|/ref|><|det|>[[117, 304, 875, 342]]<|/det|>
+I performed the requested EMSA experiments, included them as a new Supplementary Fig.7 and referred them in the text (please see the yellow marking in the PDF file for reviewers).  
+
+<|ref|>text<|/ref|><|det|>[[117, 353, 266, 378]]<|/det|>
+Sincerely, Dr.Daria Onichtchouk.  
+
+<|ref|>text<|/ref|><|det|>[[118, 400, 373, 414]]<|/det|>
+Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[117, 425, 570, 439]]<|/det|>
+All of my concerns regarding this manuscript have been addressed.
+
+<--- Page Split --->

peer_reviews/1292826cb041fd67ca20b49f8643de61f387627c09dfbb550f24e651c23d0208/supplementary_0_Transparent Peer Review file/images_list.json

@@ -0,0 +1 @@
+[]

peer_reviews/1292826cb041fd67ca20b49f8643de61f387627c09dfbb550f24e651c23d0208/supplementary_0_Transparent Peer Review file/supplementary_0_Transparent Peer Review file.mmd

@@ -0,0 +1,451 @@
+
+# nature portfolio  
+
+Peer Review File  
+
+# The role of cytochrome bc1 inhibitors in future tuberculosis treatment regimens  
+
+Corresponding Author: Dr Clara Aguilar Pérez  
+
+This file contains all reviewer reports in order by version, followed by all author rebuttals in order by version.  
+
+Version 0:  
+
+Reviewer comments:  
+
+Reviewer #1  
+
+(Remarks to the Author)  
+
+The manuscript by Perez and colleagues present data on the effects of cytochrome bc1 inhibitors and their potential efficacy in drug regimens against multidrug resistant TB as well as its role in reducing the duration of chemotherapy in drug susceptible TB. The work is important and timely considered the burden of drug- resistant TB and the prolonged treatment protocols for both drug resistant and susceptible cases of TB. The experimental approach is solid and data presented are robust to support their hypotheses.  
+
+One of the main interesting points in the manuscript is the fact that cytochrome bc1 inhibitors are more potent against clinical isolates than the lab strains. A further investigation into this would have increased the impact of the work as well as our understanding of the mechanisms of action of these inhibitors. Especially as these are claimed to be validated inhibitors, it would also be nice to see more of their characterisation and target engagement in vitro.  
+
+In addition, would this variability on strain efficacy mean that there are potential other targets being inhibited by the same compounds or is it the secondary effects of inhibiting the cytochrome bc1 and the respiratory chain of Mtb?  
+
+For the present study, authors can expand the discussion regarding the efficacy of their compound in the background of bd and how this could affect the outcome. More importantly, the authors can use these compounds to further interrogate this hypothesis. In addition, having another look at their data, I could see that if you calculate the MBC/MIC ratio for the reference lab strain (H37Rv) and the cytochrome deleted mutant (H37Rv_CytBd- KO) on supplementary table S2 one can see that their inhibitor is bacteriostatic on H37Rv whereas it is bactericidal on cytochrome bd deleted mutant (H37Rv_CytBd- KO). That further adds to the argument that the expression levels of cytochrome bd is an important factor in their compound efficacy.  
+
+Furthermore, the authors did not include if these compounds are active against MDR and XDR Mtb strains. Taking into account that the main use of these inhibitors would be towards a drug regimen for treating MDR- TB, it would be nice to see that the compounds maintain their activity against a panel of multiple drug- resistant strains including MDR and XDR clinical isolates.  
+
+Obviously a series of experiments where one looks at the levels of expression of these genes under treatment or not in a series of lab and clinical strains of Mtb would shed light on this question. But I don't think these experiments need to be part of this work.  
+
+Lastly, in the Figure S3, I cannot see the control line as well as the figure legend need a bit more explanation.  
+
+## Reviewer #2  
+
+(Remarks to the Author) The authors have done an excellent job of identifying cytochrome bc1 as an important component of the Mtb electron transport chain and a potentially sterilizing target in for TB regimen development. Using a relapsing mouse model, they have attempted to demonstrate that cytochrome bc1 inhibitors could be effective partner drugs in TB regimen development that could enhance sterilization. They report on several novel regimens
+
+<--- Page Split --->
+
+as examples of this role for bc1 inhibitors in regimens for both multidrug- resistant TB (MDR- TB) and drug- sensitive TB (DSTB), where cytochrome bc1 inhibitors could contribute to sterilization and treatment shortening. importantly, they have shown that clinical isolates exhibit heightened susceptibility to cytochrome bc1 inhibitors compared to laboratory- adapted strains, further supporting their potential usefulness in a clinical setting even as their contribution may be low or undetectable against a lab adapted strain. They take these findings as confirmation that cytochrome bc1 inhibitors have significant potential to improve TB treatment outcomes and highlight the need for further studies to evaluate their clinical contribution to novel treatment regimens.  
+
+What are the noteworthy results? The authors have done an excellent job of identifying and advancing highly potent inhibitors against their chosen target, cytochrome bc1. There is enough evidence to demonstrate that this is a sound choice of target, especially with the excellent success of bedaquiline, the flagship ATP synthase developed by the same group that has literally changed the paradigm in TB treatment shortening and highlighted the electron transport chain as a highly valuable and sterilizing target, at the same level or better than the rifamycin of prior years. It is therefore sound and admirable to hit this pathway as hard as possible, at multiple points and potentially achieve better cure. The problem seems to be that the interplay between cytochrome bc and cytochrome bd seems complicated and makes it difficult to determine with available data, including the data reported here, if just inhibiting bc, but not bd will be enough to realize the expected benefits. It seems that high expression of bd is able to reverse or nullify the effects of inhibiting bc. This is potentially the reason why there is strain difference between lab adapted versus clinical isolates. Further proof is gleaned from the excellent performance of Q203, the lead asset in this target space, versus mycobacterial species that do not express bd like M. leprae and M. ulcerans. This leaves the potential for a bc inhibitor uncertain against tuberculosis. Having said that, the team has identified a tool compound that together with Q203 will be useful in further investigating this target  
+
+Will the work be of significance to the field and related fields? While the actual developability of this asset in future TB drug regimens is in doubt until more work is done to fully appreciate contribution against all TB strains, publication of this work will be of great significance to the field and specifically to the grand idea of targeting the electron transport chain of Mtb at multiple nodes to achieve better cure for TB. Furthermore, the compound reported here could also be expanded to the NTM space as is being done for Q203 which is currently being evaluated against leprosy and buruli ulcer, arguably important and neglected diseases as well. But for use in tuberculosis, more work needs to be done to understand the importance of targeting bc1 in a background with high expression of bd. And to determine if the site of infection and location in a lesion affects the usefulness of this asset. And if the mouse model as used today is the best model for this particular asset  
+
+How does it compare to the established literature? The work reported here is of a high level, similar to other novel drug development programs. Importantly, this was not just a random shot in the dark, instead it was a highly impressive hypothesis driven mission that actually delivered a high- quality compound that is available for further evaluation to prove the hypothesis and potentially deliver a developable asset. It could be argued that in some of the regimen development, drug combination studies in mice, a more factorial approach that builds from 2, to 3 and then perhaps four drugs, with the ability to clearly demonstrate the sterilizing contribution of the asset being evaluated could have been more rigorous than reported here. This leaves this reviewer unable in some cases to confirm that the J bc1 inhibitor contributed anything at all in some of the reported combinations. This could be cleaned up on a re- submission, or in future studies that would attempt to demonstrate the full potential of this asset.  
+
+If the work is not original, please provide relevant references. The work is original and very well done  
+
+Does the work support the conclusions and claims, or is additional evidence needed? The work supports some of the claims, but as stated above, the usefulness of developing a bc1 inhibitors and that inhibitor demonstrating a real contribution to treatment shortening against a tuberculosis infection still needs more work.  
+
+Are there any flaws in the data analysis, interpretation and conclusions? Target validation for bc1 in a highly expressed bd background needs special attention. Contribution of the asset to significant treatment shortening in a developable regimen could be improved  
+
+Do these prohibit publication or require revision? I recommend publication of this work with a more substantive discussion of the strain difference and usefulness of bc1 inhibitors vs bd expression and a better demonstration of contribution to treatment shortening  
+
+Is the methodology sound? The methods used are sound except for the identified shortcomings  
+
+Does the work meet the expected standards in your field? The work is high quality from a highly experienced group working in a target pathway with which they are clearly experts. The regimen development work, especially with this challenging target, relative to bd expression could be improved  
+
+Is there enough detail provided in the methods for the work to be reproduced? Yes  
+
+Recommendation: Accept for publication with some revision of the highlighted aspects, above  
+
+Reviewer #3  
+
+(Remarks to the Author)
+
+<--- Page Split --->
+
+## General Comments  
+
+The manuscript presents valuable findings on the role of cytochrome bc1 inhibitors in tuberculosis treatment. However, it currently lacks a clear and well- organized structure, which makes it challenging to follow. To be considered suitable for publication, we strongly recommend rewriting the whole manuscript to include distinct and complete sections—specifically, Introduction, Materials & Methods, Results, Discussion, and Conclusion. The current format hinders readability and clarity and omits critical information necessary to fully assess the study's findings. Additionally, the manuscript would benefit from enhanced experimental justification and a more thorough discussion of the results.  
+
+The experimental design, including details on mice strains and compound combinations and its rationale, should be consistently presented across all sections (Introduction, Results, Materials & Methods) and thoroughly discussed. Introduction  
+
+The manuscript provides a well- articulated description of the problem and effectively highlights the potential of the electron transport chain inhibitors in tuberculosis treatment. However, we miss a proper introduction and a clear justification of the two new drug candidates (JNJ- 2901 and JNJ- 4052) that the authors test (according to the figures), and the differences and their potential benefits when compared to the cytochrome bc1 inhibitor Q203 (Telacebec) already being evaluated in clinical trials. This would provide proof of the manuscript's novelty.  
+
+Hypothesis, aim and objectives are not well presented, and clearly insufficiently described. Moreover, supplementary information and figures should not be included in the Introduction but rather integrated and discussed in the results section. Results  
+
+The results demonstrate promising data on JNJ- 2901's efficacy in reducing bacterial burden and preventing relapse in mice models. The findings suggest a potential shortening of TB treatment regimens, particularly with BPaCJ, which showed the highest efficacy and lowest relapse rates compared to BPaL, BPaM, BPaC, and BPaJ.  
+
+However, we have major concerns regarding the description and structure of this section:  
+
+- Global experimental plan should first be presented to help understand what the authors did, and then each set of results should be clearly introduced with a summary of the main conclusion before presenting the data of each experimental study. 
+- The experimental design (mice strain, infection mode, treatment duration, and rationale for each study) and must be clearly stated at the beginning of each results subsection. We here wanted to point out that the naming convention of studies (e.g., Study A, B, C, etc.) does not provide enough information to the reader. We suggest their replacement throughout the manuscript with descriptive terms reflecting the experimental model. For example, instead of "In an additional study (Study C)," we suggest "We additionally investigated drug combination efficacy and relapse in intravenously infected mice." 
+- Why a second compound – JNJ-4052- is tested in Study F? Why to test another candidate?  
+
+Discussion  
+
+Although a very small discussion comments are integrated into the manuscript (e.g., L136- 143, L156- 161, L171- 177), a dedicated Discussion section is necessary. We suggest moving the relevant lines into a standalone Discussion section and expanding on several key topics:  
+
+- Comparison of JNJ-2901 and JNJ-4052 with Telacebec (Q203): The manuscript should provide a comparative analysis of the new candidates with existing inhibitors in terms of efficacy, safety, and pharmacokinetics as well as treatment efficacy and relapsing rates. What is the novelty? Why these compounds are better? Why they test 2 different compounds? And if everything is done with JNJ-2901, why to include JNJ-4052 in one of the studies?  
+
+- Differences in mice models and infection routes: The study utilizes different infection models, and these differences should be considered to discuss the results. In addition, the authors should also discuss the gender bias in the experiments (only female mice are used) and how this could impact clinical translation.  
+
+- Comparison between clinical isolates and wild-type (WT) strains: Authors should also discuss the observed differences in drug efficacy between clinical isolates and WT strains. The reasons behind these differences need to be explained and appropriately discussed.  
+
+- Experimental limitations: The mention of limitations in L119 is too vague. The authors should explicitly outline the limitations of their experimental design.  
+
+- The potential of cytochrome bc1 inhibitors: The claim that these inhibitors could significantly impact TB treatment is overstated for the level of preclinical data presented. Further discussion on the limitations of the preclinical findings hereby presented and next steps for clinical translation is needed, including challenges, barriers and further work to be done before undergoing clinical evaluation.  
+
+## Materials and Methods  
+
+The Materials & Methods section is well- detailed and allows for experimental reproducibility. However, there we have some comments:  
+
+- As in the Results section, avoid referring only to different experiments as Study A, B, etc. Instead, add descriptive headings such as "Intranasal Mice Challenge", "Intravenous Mice Challenge", or "High-Dose Aerosol Challenge."  
+
+- Each experimental model should consistently include the following information: Mouse strain, sex, and age; inoculum  
+
+preparation, infection route, and dose; treatment duration and euthanasia time points.  
+
+- Authors should provide in the ethical statement the end-point criteria.  
+
+- Supplementary tables and results should not be included in the Methods section but should be properly cited in the Results section.  
+
+- There are several experimental methods referenced in the supplementary material but not appearing in the main text (e.g. PK and tolerability in mouse, ADME assays, HepG2 cytotoxicity assay, mitochondrial toxicity assay Glu/Gal, Ames II Mutagenicity assay, mutant isolation and WGS). These should be briefly mentioned in the main text, such as: "Both JNJ-2901 and JNJ-4052 were evaluated for ADME and toxicity. All parameters supported further use of these compounds in subsequent experiments (Supplementary Table S2)".  
+
+Additionally, we suggest to address the following minor comments:  
+
+- L62: we would suggest classifying tuberculosis as a global pandemic. 
+- L79: MoA abbreviation is not needed.
+
+<--- Page Split --->
+
+- L87: the word "validated" to describe the inhibitor compound JNJ-2901 is confusing as the presented manuscript seems to be the validation of the compound as an effective treatment adjuvant. If this is not the case, please provide the reference where the compound is validated for its inhibition of the cytochrome bc1.  
+
+- L91-96: we suggest to relocate this in the line 76, as the TB-PRACTECAL trial results are a good justification of linezolid-associated adverse effects when treating MDR-TB.  
+
+- L116-122: please contextualize the results from this part.  
+
+- L127: move the figure citation to line 133. After the results are explained.  
+
+- L146: please provide the justification of why using Telacobec in this study and not JNJ-2901 and provide rationale about the drug regimen used for this experimental design.  
+
+- L164: as far as we understood, JNJ-2901 and Telacobec were also assessed for its bactericidal activity against clinical isolates in comparison to the H37Rv strain (Figure 2 B,C,D and E). Please provide a proper description of the compound used for this experiment and further discussion of this results in the main text.  
+
+- Figure 1: If the mice strain and experimental design are the same, consider re-structure Figure 1B and 1C in a unique figure.  
+
+- Figure 2: Please, consider a figure rearrangement. We suggest dividing Figure 2 in two. Section A on one side (resulting in Figure 2) and sections B, C, D, E and F on the other (resulting in Figure 3 A, B, C, D and E). Since the results represented refer to different experimental questions and require a results section for each one.  
+
+## Final Recommendation  
+
+This study presents valuable preclinical data on cytochrome bc1 inhibitors in TB treatment. However, we have significant concerns regarding the manuscript as it currently stands. It lacks a proper structure, clarity, and experimental justification, and the discussion is entirely insufficient. The manuscript needs to be totally rewritten to address these issues, ensuring that it achieves the required readability, scientific rigor, and overall impact to be considered suitable for publication.  
+
+## Reviewer #4  
+
+(Remarks to the Author)  
+
+I co- reviewed this manuscript with one of the reviewers who provided the listed reports. This is part of the Nature Communications initiative to facilitate training in peer review and to provide appropriate recognition for Early Career Researchers who co- review manuscripts.  
+
+Version 1:  
+
+Reviewer comments:  
+
+Reviewer #1  
+
+(Remarks to the Author)  
+
+The revised manuscript of Clara Aguilar Pérez and colleagues has addressed all of my previous concerns and I am happy to see some of the recommendation on the revised discussion section as well as extended new data. The study represents an important progress on inhibition of respiration machinery in Mtb and has potential for the development of future therapeutics.  
+
+Reviewer #3  
+
+(Remarks to the Author)  
+
+We would like to thank the authors for thoughtfully addressing our major suggestions regarding the manuscript. Our feedback primarily focused on improving its structure, clarity, and the discussion surrounding the experimental procedures and results. In particular, the authors have now effectively justified the differences observed when treating laboratory strains with different cytochrome bc1 inhibitors versus clinical strains. With these revisions, the manuscript now clearly highlights the significance of this investigation in advancing our understanding of tuberculosis treatment strategies.  
+
+This work underscores the potential of bc1 inhibitors as viable alternatives to fluoroquinolones and linezolid, which are often associated with adverse effects in current regimens for drug- resistant tuberculosis. Moreover, it contributes to the possibility of treatment shortening for drug- sensitive tuberculosis, with added potential against clinical strains. The experiments were performed using different, well- justified preclinical mouse models, which strengthens the translational relevance of the findings and supports further investigation of this promising asset. Based on these improvements and the scientific merit of the study, we believe the manuscript is now better suited for publication.  
+
+However, we would like the authors to address the following remaining issue:  
+
+Please add a paragraph on limitations to the discussion section. For example, the sample size of mice used in the experiments could be acknowledged as a limitation (even though we recognize that this is already mentioned on page 7). Additionally, regarding the authors' response that "no differences between sexes have been reported in relation to drug efficacy in TB mouse models," we would like to stress that, even if ethical requirements do not mandate the inclusion of both biological sexes, doing so is strongly advisable. There is robust evidence from both clinical and preclinical studies that biological sex influences TB epidemiology and treatment outcomes (for example, see: Idris, R., Infection 2025, https://doi.org/10.1007/s15010-024-02424-5; Dutta NK, Front Immunol, 2020, doi: 10.3389/fimmu.2020.01465; Tannenbaum,
+
+<--- Page Split --->
+
+C., Nature, 2019, https://doi.org/10.1038/s41586-019-1657-6). These differences are driven by a combination of hormonal, genetic, and immunological factors, as well as potential differences in treatment adherence and broader social determinants of health. While we acknowledge that using female mice can be more practical in terms of logistics and animal management- since males often display more aggressive behavior that can complicate experiments- there is now ample evidence showing that men and women often have different disease trajectories (not only for TB!) and responses to treatment. One persistent issue in drug development is that preclinical and clinical studies frequently fail to reflect the biological diversity found in real- world populations. Including both sexes in preclinical experiments wherever feasible would help improve the generalizability of the findings and ensure that results more accurately inform treatment strategies for all patients.  
+
+In summary, we encourage the authors to carefully consider any additional limitations of their study and clearly state them, as this will add valuable context and transparency. Once this is done, the manuscript can be endorsed for publication.  
+
+## Reviewer #4  
+
+(Remarks to the Author) I co- reviewed this manuscript with one of the reviewers who provided the listed reports. This is part of the Nature Communications initiative to facilitate training in peer review and to provide appropriate recognition for Early Career Researchers who co- review manuscripts.  
+
+Open Access This Peer Review File is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.  
+
+In cases where reviewers are anonymous, credit should be given to 'Anonymous Referee' and the source. The images or other third party material in this Peer Review File are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.  
+
+To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/
+
+<--- Page Split --->
+
+## Reviewers' comments  
+
+We thank the reviewers for providing useful comments and suggestions that have greatly improved this manuscript. We have addressed each comment below (red).  
+
+## Reviewer #1 (Remarks to the Author):  
+
+The manuscript by Perez and colleagues present data on the effects of cytochrome bc1 inhibitors and their potential efficacy in drug regimens against multidrug resistant TB as well as its role in reducing the duration of chemotherapy in drug susceptible TB. The work is important and timely considered the burden of drug- resistant TB and the prolonged treatment protocols for both drug resistant and susceptible cases of TB. The experimental approach is solid and data presented are robust to support their hypotheses.  
+
+One of the main interesting points in the manuscript is the fact that cytochrome bc1 inhibitors are more potent against clinical isolates than the lab strains. A further investigation into this would have increased the impact of the work as well as our understanding of the mechanisms of action of these inhibitors. Especially as these are claimed to be validated inhibitors, it would also be nice to see more of their characterisation and target engagement in vitro.  
+
+A second paper has recently been published which describes the characterisation, validation and target engagement of the JNJ- 2901 inhibitor (PMID:40191462). We have provided validation of JNJ- 4052, which comes from the same chemical series as JNJ- 2901, in Supplementary Tables S1- 3.  
+
+In addition, would this variability on strain efficacy mean that there are potential other targets being inhibited by the same compounds or is it the secondary effects of inhibiting the cytochrome bc1 and the respiratory chain of Mtb?  
+
+We do not believe that there are any genuine secondary targets for these compounds. Our resistance data demonstrates that a single nucleotide polymorphism in qcrB can provide high level resistance (for example T313A leads to \(>2000\) - fold resistance to JNJ- 4052; Supplementary Table S1). As discussed below, we believe the differences are most likely due to the increased expression of cytochrome bd, an alternative oxidase that can compensate for cytochrome bc1 inhibition. This is now outlined in the Discussion section (pg 9- 10).  
+
+For the present study, authors can expand the discussion regarding the efficacy of their compound in the background of bd and how this could affect the outcome. More importantly, the authors can use these compounds to further interrogate this hypothesis. In addition, having another look at their data, I could see that if you calculate the MBC/MIC ratio for the reference lab strain (H37Rv) and the cytochrome deleted mutant (H37Rv_CytBd- KO) on supplementary table S2 one can see that their inhibitor is bacteriostatic on H37Rv whereas it is bactericidal on cytochrome bd deleted mutant (H37Rv_CytBd- KO). That further adds to the argument that the expression levels of cytochrome bd is an important factor in their compound efficacy.  
+
+We agree that this is an important point and have now provided discussion of the difference in bactericidality between H37Rv and H37Rv_CytBd- KO strains in the Discussion section (pg 9- 10).  
+
+Furthermore, the authors did not include if these compounds are active against MDR and XDR Mtb strains. Taking into account that the main use of these inhibitors would be towards a drug regimen for treating MDR- TB, it would be nice to see that the compounds maintain their activity against a panel of multiple drug- resistant strains including MDR and XDR clinical isolates.
+
+<--- Page Split --->
+
+Activity of JNJ- 2901 against a panel of 18 MDR- TB strains has now been published showing that the inhibitor is more active compared to lab- adapted strains (PMID:40191462).  
+
+Obviously a series of experiments where one looks at the levels of expression of these genes under treatment or not in a series of lab and clinical strains of Mtb would shed light on this question. But I don't think these experiments need to be part of this work.  
+
+Lastly, in the Figure S3, I cannot see the control line as well as the figure legend need a bit more explanation.  
+
+We have updated the figure legend to indicate that moxifloxacin is hidden by BDQ (Supplementary Figure S3). We have added more detail of the experimental conditions in the figure legend.  
+
+## Reviewer #2 (Remarks to the Author):  
+
+The authors have done an excellent job of identifying cytochrome bc1 as an important component of the Mtb electron transport chain and a potentially sterilizing target in for TB regimen development. Using a relapsing mouse model, they have attempted to demonstrate that cytochrome bc1 inhibitors could be effective partner drugs in TB regimen development that could enhance sterilization. They report on several novel regimens as examples of this role for bc1 inhibitors in regimens for both multidrug- resistant TB (MDR- TB) and drug- sensitive TB (DS- TB), where cytochrome bc1 inhibitors could contribute to sterilization and treatment shortening. importantly, they have shown that clinical isolates exhibit heightened susceptibility to cytochrome bc1 inhibitors compared to laboratory- adapted strains, further supporting their potential usefulness in a clinical setting even as their contribution may be low or undetectable against a lab adapted strain. They take these findings as confirmation that cytochrome bc1 inhibitors have significant potential to improve TB treatment outcomes and highlight the need for further studies to evaluate their clinical contribution to novel treatment regimens.  
+
+What are the noteworthy results? The authors have done an excellent job of identifying and advancing highly potent inhibitors against their chosen target, cytochrome bc1. There is enough evidence to demonstrate that this is a sound choice of target, especially with the excellent success of bedaquiline, the flagship ATP synthase developed by the same group that has literally changed the paradigm in TB treatment shortening and highlighted the electron transport chain as a highly valuable and sterilizing target, at the same level or better than the rifamycins of prior years. It is therefore sound and admirable to hit this pathway as hard as possible, at multiple points and potentially achieve better cure. The problem seems to be that the interplay between cytochrome bc and cytochrome bd seems complicated and makes it difficult to determine with available data, including the data reported here, if just inhibiting bc, but not bd will be enough to realize the expected benefits. It seems that high expression of bd is able to reverse or nullify the effects of inhibiting bc. This is potentially the reason why there is strain difference between lab adapted versus clinical isolates. Further proof is gleaned from the excellent performance of Q203, the lead asset in this target space, versus mycobacterial species that do not express bd like M. leprae and M. ulcerans. This leaves the potential for a bc inhibitor uncertain against tuberculosis. Having said that, the team has identified a tool compound that together with Q203 will be useful in further investigating this target  
+
+Will the work be of significance to the field and related fields? While the actual developability of this asset in future TB drug regimens is in doubt until more work is done to fully appreciate contribution
+
+<--- Page Split --->
+
+against all TB strains, publication of this work will be of great significance to the field and specifically to the grand idea of targeting the electron transport chain of Mtb at multiple nodes to achieve better cure for TB. Furthermore, the compound reported here could also be expanded to the NTM space as is being done for Q203 which is currently being evaluated against leprosy and buruli ulcer, arguably important and neglected diseases as well. But for use in tuberculosis, more work needs to be done to understand the importance of targeting bc1 in a background with high expression of bd. And to determine if the site of infection and location in a lesion affects the usefulness of this asset. And if the mouse model as used today is the best model for this particular asset  
+
+How does it compare to the established literature? The work reported here is of a high level, similar to other novel drug development programs. Importantly, this was not just a random shot in the dark, instead it was a highly impressive hypothesis driven mission that actually delivered a high- quality compound that is available for further evaluation to prove the hypothesis and potentially deliver a developable asset. It could be argued that in some of the regimen development, drug combination studies in mice, a more factorial approach that builds from 2, to 3 and then perhaps four drugs, with the ability to clearly demonstrate the sterilizing contribution of the asset being evaluated could have been more rigorous than reported here. This leaves this reviewer unable in some cases to confirm that the J bc1 inhibitor contributed anything at all in some of the reported combinations. This could be cleaned up on a re- submission, or in future studies that would attempt to demonstrate the full potential of this asset.  
+
+A more stepwise factorial approach (building from two- to three- and then four- drug combinations) could have further clarified the sterilising contribution of a cytochrome \(bc_{1}\) inhibitor. However, the aim of these studies was not to develop an entirely new regimen but rather to replace moxifloxacin and linezolid with a potentially superior compound while maintaining the existing backbone. Therefore, we prioritised evaluating the \(bc_{1}\) inhibitor in combination with a clinically relevant regimen to reflect a realistic therapeutic setting.  
+
+Additionally, we were mindful of the ethical considerations regarding animal use. Expanding the study design to include a larger number of combinations/conditions would have significantly increased the number of animals required.  
+
+In several of our studies, the addition of a cytochrome \(bc_{1}\) inhibitor led to a statistical improvement in relapse rates (e.g. Figure 2; CZ vs CZT). Nevertheless, we agree that future investigations could employ a more granular factorial design to better define the specific contribution of the \(bc_{1}\) inhibitor. These follow- up studies would help further demonstrate the full potential of cytochrome \(bc_{1}\) inhibitors and strengthen its case for clinical development.  
+
+If the work is not original, please provide relevant references. The work is original and very well done  
+
+Does the work support the conclusions and claims, or is additional evidence needed? The work supports some of the claims, but as stated above, the usefulness of developing a bc1 inhibitors and that inhibitor demonstrating a real contribution to treatment shortening against a tuberculosis infection still needs more work.  
+
+While we acknowledge that additional studies are needed to fully establish the potential of cytochrome \(bc_{1}\) inhibitors to shorten treatment duration, particularly in the context of multidrug- resistant tuberculosis (MDR- TB), the data presented here clearly demonstrate that the inclusion of a cytochrome \(bc_{1}\) inhibitor reduces the treatment time required. For example, the addition of telacebec to CZ reduces the treatment by at least 2 months (Figure 2). This is the starting point for further work to investigate this in more detail.
+
+<--- Page Split --->
+
+Are there any flaws in the data analysis, interpretation and conclusions? Target validation for bc1 in a highly expressed bd background needs special attention.  
+
+We have recently published the validation of JNJ- 2901 (PMID: 40191462) and have expanded our discussion on bd expression (pg 9- 10).  
+
+Contribution of the asset to significant treatment shortening in a developable regimen could be improved  
+
+As discussed above, we view this work as an important starting point for investigating the potential contribution of cytochrome \(bc_{1}\) inhibitors to future treatment- shortening regimens. We agree that further work is required beyond the current study. However, one challenge in addressing this more definitively is the lack of consensus on what constitutes a 'developable' regimen, as different organisations often have varying perspectives on which drugs should be prioritised; clofazimine being a notable example here.  
+
+Do these prohibit publication or require revision? I recommend publication of this work with a more substantive discussion of the strain difference and usefulness of bc1 inhibitors vs bd expression and a better demonstration of contribution to treatment shortening  
+
+We have now re- formatted the manuscript and expanded the discussion on the clinical isolate differences we observed.  
+
+Is the methodology sound? The methods used are sound except for the identified shortcomings  
+
+Does the work meet the expected standards in your field? The work is high quality from a highly experienced group working in a target pathway with which they are clearly experts. The regimen development work, especially with this challenging target, relative to bd expression could be improved  
+
+Is there enough detail provided in the methods for the work to be reproduced? Yes Recommendation: Accept for publication with some revision of the highlighted aspects, above  
+
+## Reviewer #3 and #4 (Remarks to the Author):  
+
+General Comments  
+
+The manuscript presents valuable findings on the role of cytochrome bc1 inhibitors in tuberculosis treatment. However, it currently lacks a clear and well- organized structure, which makes it challenging to follow. To be considered suitable for publication, we strongly recommend rewriting the whole manuscript to include distinct and complete sections—specifically, Introduction, Materials & Methods, Results, Discussion, and Conclusion. The current format hinders readability and clarity and omits critical information necessary to fully assess the study's findings. Additionally, the manuscript would benefit from enhanced experimental justification and a more thorough discussion of the results.  
+
+We have now re- formatted the manuscript for Nature Communications as suggested and expanded on the Discussion section.  
+
+The experimental design, including details on mice strains and compound combinations and its
+
+<--- Page Split --->
+
+rationale, should be consistently presented across all sections (Introduction, Results, Materials & Methods) and thoroughly discussed.  
+
+## Introduction  
+
+The manuscript provides a well- articulated description of the problem and effectively highlights the potential of the electron transport chain inhibitors in tuberculosis treatment. However, we miss a proper introduction and a clear justification of the two new drug candidates (JNJ- 2901 and JNJ- 4052) that the authors test (according to the figures), and the differences and their potential benefits when compared to the cytochrome bc1 inhibitor Q203 (Telacebec) already being evaluated in clinical trials. This would provide proof of the manuscript's novelty.  
+
+The primary focus of this manuscript is to understand the contribution of a cytochrome bc1- targeting inhibitor to TB treatment regimens, rather than specifically focusing on JNJ- 2091 or suggesting it is a superior molecule. We consider all the compounds discussed in this work as tool compounds used to increase our understanding of how this mode of action could complement existing drug regimens. In parallel, we have published a separate validation of JNJ- 2901 (PMID:40191462) and have added additional justification here (pg 5- 6).  
+
+Hypothesis, aim and objectives are not well presented, and clearly insufficiently described.  
+
+We have clarified and expanded the objectives of each section as requested.  
+
+Moreover, supplementary information and figures should not be included in the Introduction but rather integrated and discussed in the results section.  
+
+Reference to the figures and supplementary information has been removed from the introduction following re- formatting.  
+
+## Results  
+
+The results demonstrate promising data on JNJ- 2901's efficacy in reducing bacterial burden and preventing relapse in mice models. The findings suggest a potential shortening of TB treatment regimens, particularly with BPaCJ, which showed the highest efficacy and lowest relapse rates compared to BPaL, BPaM, BPaC, and BPaJ.  
+
+However, we have major concerns regarding the description and structure of this section:  
+
+- Global experimental plan should first be presented to help understand what the authors did, and then each set of results should be clearly introduced with a summary of the main conclusion before presenting the data of each experimental study.  
+
+We agree, and have now divided this section more clearly with Figures 1b- c now discussed sequentially. An overview of the first set of experiments is provided in Figure 1a.  
+
+- The experimental design (mice strain, infection mode, treatment duration, and rationale for each study) and must be clearly stated at the beginning of each results subsection. We here wanted to point out that the naming convention of studies (e.g., Study A, B, C, etc.) does not provide enough information to the reader. We suggest their replacement throughout the manuscript with descriptive terms reflecting the experimental model. For example, instead of “In an additional study (Study C),” we suggest “We additionally investigated drug combination efficacy and relapse in intravenously infected mice.”
+
+<--- Page Split --->
+
+We respectfully disagree, as the experimental design is indicated in the Materials and Methods section. We do not believe it is necessary to repeat this information in the Results section or figure legends. We have named them A, B, C etc. because the studies were performed by different institutions rather than because they are different type of studies.  
+
+- Why a second compound – JNJ-4052- is tested in Study F? Why to test another candidate?  
+
+JNJ- 2901 and JNJ- 4052 belong to the same chemical series and exhibit very similar activity and safety profiles. Our cytochrome \(bc_{1}\) drug discovery programme has been ongoing for 7 years, during which time we have progressed a number of lead compounds including JNJ- 4052 and JNJ- 2901. The in vivo clinical isolate study (Study F) was performed using the earlier lead, JNJ- 4052. While we acknowledge that repeating this experiment with JNJ- 2901 would have been ideal, given the comparable profiles of both compounds, we do not anticipate a change in the outcome. Therefore, we believe repeating this study would not represent a justifiable use of mice.  
+
+## Discussion  
+
+Although a very small discussion comments are integrated into the manuscript (e.g., L136- 143, L156- 161, L171- 177), a dedicated Discussion section is necessary. We suggest moving the relevant lines into a standalone Discussion section and expanding on several key topics:  
+
+- Comparison of JNJ-2901 and JNJ-4052 with Telacebec (Q203): The manuscript should provide a comparative analysis of the new candidates with existing inhibitors in terms of efficacy, safety, and pharmacokinetics as well as treatment efficacy and relapsing rates. What is the novelty? Why these compounds are better? Why they test 2 different compounds? And if everything is done with JNJ-2901, why to include JNJ-4052 in one of the studies?  
+
+The aim of the project was to investigate the role of cytochrome \(bc_{1}\) inhibitors in future treatment regimens rather than suggest that JNJ- 2901 is superior to telacebec. We selected three inhibitors with related chemical structure, mode of binding and efficacy to achieve these aims; we have made this clearer in the text (pg 5- 6).  
+
+- Differences in mice models and infection routes: The study utilizes different infection models, and these differences should be considered to discuss the results. In addition, the authors should also discuss the gender bias in the experiments (only female mice are used) and how this could impact clinical translation.  
+
+In this study, two different infection models were used: nasal (via aerosol or intranasal instillation) and intravenous. As noted in the text, we acknowledge that the intravenous model exhibits a higher severity of the disease since this infection route confers a systemic infection, whereas the nasal route mainly leads to a pulmonary infection. The intention of this study was not to characterise the treatment response at the pathology level, but rather to have a broad knowledge of the activity of cytochrome \(bc_{1}\) - containing regimens in multiple potential scenarios.  
+
+Regarding the use of only female mice, we acknowledge recommendations for including both sexes in animal studies. However, in line with our institutional policies focused on the 3Rs, we used only female mice to minimise stress and harm. Relapse studies are very long and involve repeated daily handling and dosing, which can lead to a higher risk of aggres behaviour in males. Clinical translation would take place when human exposures are known in a Phase I trial. Additionally, no differences between sex have been reported in relation to drug efficacy in TB mouse models.  
+
+- Comparison between clinical isolates and wild-type (WT) strains: Authors should also discuss the
+
+<--- Page Split --->
+
+observed differences in drug efficacy between clinical isolates and WT strains. The reasons behind these differences need to be explained and appropriately discussed.  
+
+We have now added discussion on the comparison of WT vs. clinical isolate (pg 9- 10).  
+
+- Experimental limitations: The mention of limitations in L119 is too vague. The authors should explicitly outline the limitations of their experimental design.  
+
+The main limitation of this study was the number of animals used did not provide enough statistical power to differentiate the different conditions. We have mentioned this in the text (pg 7).  
+
+- The potential of cytochrome bc1 inhibitors: The claim that these inhibitors could significantly impact TB treatment is overstated for the level of preclinical data presented. Further discussion on the limitations of the preclinical findings hereby presented and next steps for clinical translation is needed, including challenges, barriers and further work to be done before undergoing clinical evaluation.  
+
+As discussed above, we wish to make it clear that JNJ- 2901 is a tool compound that we have used to demonstrate how, inhibitors of cytochrome bc1, a novel mode of action, would interact with existing treatments. Our conclusions refer to 'cytochrome bc1 inhibitors' rather than to a specific compound in development.  
+
+## Materials and Methods  
+
+The Materials & Methods section is well- detailed and allows for experimental reproducibility. However, there we have some comments:  
+
+- As in the Results section, avoid referring only to different experiments as Study A, B, etc. Instead, add descriptive headings such as "Intranasal Mice Challenge", "Intravenous Mice Challenge", or "High-Dose Aerosol Challenge."  
+
+- Each experimental model should consistently include the following information: Mouse strain, sex, and age; inoculum preparation, infection route, and dose; treatment duration and euthanasia time points.  
+
+We respectfully disagree with this suggestion. The experimental details for each study are already included in the Methods section, and adding them to the Results and figure legends would lead to unnecessary repetition. We have intentionally used this naming strategy to clearly distinguish studies from different institutions. Furthermore, many of the models have similar experimental parameters, making it difficult to differentiate them by name alone.  
+
+- Authors should provide in the ethical statement the end-point criteria.  
+
+We have now included the humane end points for the in vivo studies (pg 18).  
+
+- Supplementary tables and results should not be included in the Methods section but should be properly cited in the Results section.  
+
+Where possible, we have discussed supplementary tables and results in the Results section, however we disagree that these elements cannot be cited in the methods.  
+
+- There are several experimental methods referenced in the supplementary material but not appearing in the main text (e.g PK and tolerability in mouse, ADME assays, HepG2 cytotoxicity assay, mitochondrial toxicity assay Glu/Gal, Ames II Mutagenicity assay, mutant isolation and WGS). These should be briefly mentioned in the main text, such as: "Both JNJ-2901 and JNJ-4052
+
+<--- Page Split --->
+
+were evaluated for ADME and toxicity. All parameters supported further use of these compounds in subsequent experiments (Supplementary Table S2)".  
+
+We have now included this information in the Results section (pg 5- 6).  
+
+Additionally, we suggest to address the following minor comments: - L62: we would suggest classifying tuberculosis as a global pandemic.  
+
+We have updated this as suggested.  
+
+- L79: MoA abbreviation is not needed.  
+
+'MoA' is used elsewhere in the manuscript.  
+
+- L87: the word "validated" to describe the inhibitor compound JNJ-2901 is confusing as the presented manuscript seems to be the validation of the compound as an effective treatment adjuvant. If this is not the case, please provide the reference where the compound is validated for its inhibition of the cytochrome bc1.  
+
+We have kept 'validated' as we can now present an additional manuscript (PMID:40191462), which outlines this validation.  
+
+- L91-96: we suggest to relocate this in the line 76, as the TB-PRACTECAL trial results are a good justification of linezolid-associated adverse effects when treating MDR-TB.  
+
+We have now moved this section to the Introduction as suggested.  
+
+- L116-122: please contextualize the results from this part.  
+
+We have expanded the results as requested.  
+
+- L127: move the figure citation to line 133. After the results are explained.  
+
+This has been moved as requested.  
+
+- L146: please provide the justification of why using Telacebec in this study and not JNJ-2901 and provide rationale about the drug regimen used for this experimental design.  
+
+This study was inspired by previous in vitro work demonstrating that the combination of bedaquiline, clofazimine, and telacebec resulted in rapid killing (PMID:27506290). We aimed to evaluate the in vivo efficacy of treatment regimens containing these compounds in combination with pyrazinamide, which has also been shown to enhance efficacy, potentially supporting treatment shortening (PMID:25622149). To allow comparison with previous work, we continued to use telacebec in this study rather than JNJ-2901. We have now clarified this justification in the results (pg 7) and discussion (pg 9).  
+
+- L164: as far as we understood, JNJ-2901 and Telacebec were also assessed for its bactericidal activity against clinical isolates in comparison to the H37Rv strain (Figure 2 B,C,D and E). Please provide a proper description of the compound used for this experiment and further discussion of this results in the main text.  
+
+At the time that this experiment was initiated, JNJ- 4052 was our lead molecule from our cytochrome bc1 drug discovery programme. This was the main reason it was chosen for the in vivo study outlined in Figure 3e. This compound provides a proof-of-principle for inhibition of cytochrome bc1, with other compounds with this MoA expected to have the same response as suggested by Figure 3b-d.
+
+<--- Page Split --->
+
+- Figure 1: If the mice strain and experimental design are the same, consider re-structure Figure 1B and 1C in a unique figure.  
+
+We would prefer to keep these studies separate, enabling the reader to discriminate between the individual experiments.  
+
+- Figure 2: Please, consider a figure rearrangement. We suggest dividing Figure 2 in two. Section A on one side (resulting in Figure 2) and sections B, C, D, E and F on the other (resulting in Figure 3 A, B, C, D and E). Since the results represented refer to different experimental questions and require a results section for each one.  
+
+We agree, and have now split Figure 2 into two new figures as suggested.  
+
+## Final Recommendation  
+
+This study presents valuable preclinical data on cytochrome bc1 inhibitors in TB treatment. However, we have significant concerns regarding the manuscript as it currently stands. It lacks a proper structure, clarity, and experimental justification, and the discussion is entirely insufficient. The manuscript needs to be totally rewritten to address these issues, ensuring that it achieves the required readability, scientific rigor, and overall impact to be considered suitable for publication.
+
+<--- Page Split --->
+
+## Response to reviewer's comments  
+
+We have addressed the final reviewer's comment in red below. We once again thank the reviewer's for their valuable comments.  
+
+Reviewer #1 (Remarks to the Author):  
+
+The revised manuscript of Clara Aguilar Pérez and colleagues has addressed all of my previous concerns and I am happy to see some of the recommendation on the revised discussion section as well as extended new data. The study represents an important progress on inhibition of respiration machinery in Mtb and has potential for the development of future therapeutics.  
+
+Reviewer #3 (Remarks to the Author):  
+
+We would like to thank the authors for thoughtfully addressing our major suggestions regarding the manuscript. Our feedback primarily focused on improving its structure, clarity, and the discussion surrounding the experimental procedures and results. In particular, the authors have now effectively justified the differences observed when treating laboratory strains with different cytochrome bc1 inhibitors versus clinical strains. With these revisions, the manuscript now clearly highlights the significance of this investigation in advancing our understanding of tuberculosis treatment strategies.  
+
+This work underscores the potential of bc1 inhibitors as viable alternatives to fluoroquinolones and linezolid, which are often associated with adverse effects in current regimens for drug- resistant tuberculosis. Moreover, it contributes to the possibility of treatment shortening for drug- sensitive tuberculosis, with added potential against clinical strains. The experiments were performed using different, well- justified preclinical mouse models, which strengthens the translational relevance of the findings and supports further investigation of this promising asset. Based on these improvements and the scientific merit of the study, we believe the manuscript is now better suited for publication.  
+
+However, we would like the authors to address the following remaining issue:  
+
+Please add a paragraph on limitations to the discussion section. For example, the sample size of mice used in the experiments could be acknowledged as a limitation (even though we recognize that this is already mentioned on page 7). Additionally, regarding the authors' response that "no differences between sexes have been reported in relation to drug efficacy in TB mouse models," we would like to stress that, even if ethical requirements do not mandate the inclusion of both biological sexes, doing so is strongly advisable. There is robust evidence from both clinical and preclinical studies that biological sex influences TB epidemiology and treatment outcomes(for example, see: Idris, R., Infection 2025, https://doi.org/10.1007/s15010-024-02424-5; Dutta NK, Front Immunol, 2020, doi: 10.3389/fimmu.2020.01465; Tannenbaum, C., Nature, 2019, https://doi.org/10.1038/s41586-019-1657-6). These differences are driven by a combination of hormonal, genetic, and immunological factors, as well as potential differences in treatment adherence and broader social determinants of health. While we acknowledge that using female mice can be more
+
+<--- Page Split --->
+
+practical in terms of logistics and animal management- since males often display more aggressive behavior that can complicate experiments- there is now ample evidence showing that men and women often have different disease trajectories (not only for TB!) and responses to treatment. One persistent issue in drug development is that preclinical and clinical studies frequently fail to reflect the biological diversity found in real- world populations. Including both sexes in preclinical experiments wherever feasible would help improve the generalizability of the findings and ensure that results more accurately inform treatment strategies for all patients.  
+
+We thank the reviewer for their helpful suggestion. We have now added a paragraph to the discussion (pg. 10) outlining the limitations of our study and lessons for future work. Specifically, we address the use of female mice, as suggested, and also note the following points for improvement of future studies: (i) increasing statistical power by using larger animal group sizes, (ii) incorporating monotherapy arms to enable regimen deconvolution, and (iii) including clinical isolates to better reflect the diversity and relevance of circulating M. tuberculosis strains.  
+
+In summary, we encourage the authors to carefully consider any additional limitations of their study and clearly state them, as this will add valuable context and transparency. Once this is done, the manuscript can be endorsed for publication.  
+
+Reviewer #4 (Remarks to the Author):  
+
+I co- reviewed this manuscript with one of the reviewers who provided the listed reports. This is part of the Nature Communications initiative to facilitate training in peer review and to provide appropriate recognition for Early Career Researchers who co- review manuscripts.
+
+<--- Page Split --->

peer_reviews/1292826cb041fd67ca20b49f8643de61f387627c09dfbb550f24e651c23d0208/supplementary_0_Transparent Peer Review file/supplementary_0_Transparent Peer Review file_det.mmd

@@ -0,0 +1,643 @@
+<|ref|>title<|/ref|><|det|>[[72, 53, 295, 80]]<|/det|>
+# nature portfolio  
+
+<|ref|>text<|/ref|><|det|>[[74, 96, 296, 118]]<|/det|>
+Peer Review File  
+
+<|ref|>title<|/ref|><|det|>[[72, 161, 896, 210]]<|/det|>
+# The role of cytochrome bc1 inhibitors in future tuberculosis treatment regimens  
+
+<|ref|>text<|/ref|><|det|>[[73, 224, 468, 241]]<|/det|>
+Corresponding Author: Dr Clara Aguilar Pérez  
+
+<|ref|>text<|/ref|><|det|>[[72, 274, 864, 289]]<|/det|>
+This file contains all reviewer reports in order by version, followed by all author rebuttals in order by version.  
+
+<|ref|>text<|/ref|><|det|>[[73, 326, 144, 340]]<|/det|>
+Version 0:  
+
+<|ref|>text<|/ref|><|det|>[[73, 353, 219, 367]]<|/det|>
+Reviewer comments:  
+
+<|ref|>text<|/ref|><|det|>[[73, 379, 160, 393]]<|/det|>
+Reviewer #1  
+
+<|ref|>text<|/ref|><|det|>[[73, 405, 238, 418]]<|/det|>
+(Remarks to the Author)  
+
+<|ref|>text<|/ref|><|det|>[[72, 418, 919, 485]]<|/det|>
+The manuscript by Perez and colleagues present data on the effects of cytochrome bc1 inhibitors and their potential efficacy in drug regimens against multidrug resistant TB as well as its role in reducing the duration of chemotherapy in drug susceptible TB. The work is important and timely considered the burden of drug- resistant TB and the prolonged treatment protocols for both drug resistant and susceptible cases of TB. The experimental approach is solid and data presented are robust to support their hypotheses.  
+
+<|ref|>text<|/ref|><|det|>[[72, 496, 920, 550]]<|/det|>
+One of the main interesting points in the manuscript is the fact that cytochrome bc1 inhibitors are more potent against clinical isolates than the lab strains. A further investigation into this would have increased the impact of the work as well as our understanding of the mechanisms of action of these inhibitors. Especially as these are claimed to be validated inhibitors, it would also be nice to see more of their characterisation and target engagement in vitro.  
+
+<|ref|>text<|/ref|><|det|>[[72, 561, 899, 589]]<|/det|>
+In addition, would this variability on strain efficacy mean that there are potential other targets being inhibited by the same compounds or is it the secondary effects of inhibiting the cytochrome bc1 and the respiratory chain of Mtb?  
+
+<|ref|>text<|/ref|><|det|>[[72, 600, 923, 680]]<|/det|>
+For the present study, authors can expand the discussion regarding the efficacy of their compound in the background of bd and how this could affect the outcome. More importantly, the authors can use these compounds to further interrogate this hypothesis. In addition, having another look at their data, I could see that if you calculate the MBC/MIC ratio for the reference lab strain (H37Rv) and the cytochrome deleted mutant (H37Rv_CytBd- KO) on supplementary table S2 one can see that their inhibitor is bacteriostatic on H37Rv whereas it is bactericidal on cytochrome bd deleted mutant (H37Rv_CytBd- KO). That further adds to the argument that the expression levels of cytochrome bd is an important factor in their compound efficacy.  
+
+<|ref|>text<|/ref|><|det|>[[72, 691, 916, 745]]<|/det|>
+Furthermore, the authors did not include if these compounds are active against MDR and XDR Mtb strains. Taking into account that the main use of these inhibitors would be towards a drug regimen for treating MDR- TB, it would be nice to see that the compounds maintain their activity against a panel of multiple drug- resistant strains including MDR and XDR clinical isolates.  
+
+<|ref|>text<|/ref|><|det|>[[72, 757, 916, 796]]<|/det|>
+Obviously a series of experiments where one looks at the levels of expression of these genes under treatment or not in a series of lab and clinical strains of Mtb would shed light on this question. But I don't think these experiments need to be part of this work.  
+
+<|ref|>text<|/ref|><|det|>[[70, 807, 818, 822]]<|/det|>
+Lastly, in the Figure S3, I cannot see the control line as well as the figure legend need a bit more explanation.  
+
+<|ref|>sub_title<|/ref|><|det|>[[72, 847, 161, 861]]<|/det|>
+## Reviewer #2  
+
+<|ref|>text<|/ref|><|det|>[[72, 874, 922, 940]]<|/det|>
+(Remarks to the Author) The authors have done an excellent job of identifying cytochrome bc1 as an important component of the Mtb electron transport chain and a potentially sterilizing target in for TB regimen development. Using a relapsing mouse model, they have attempted to demonstrate that cytochrome bc1 inhibitors could be effective partner drugs in TB regimen development that could enhance sterilization. They report on several novel regimens
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[72, 47, 921, 140]]<|/det|>
+as examples of this role for bc1 inhibitors in regimens for both multidrug- resistant TB (MDR- TB) and drug- sensitive TB (DSTB), where cytochrome bc1 inhibitors could contribute to sterilization and treatment shortening. importantly, they have shown that clinical isolates exhibit heightened susceptibility to cytochrome bc1 inhibitors compared to laboratory- adapted strains, further supporting their potential usefulness in a clinical setting even as their contribution may be low or undetectable against a lab adapted strain. They take these findings as confirmation that cytochrome bc1 inhibitors have significant potential to improve TB treatment outcomes and highlight the need for further studies to evaluate their clinical contribution to novel treatment regimens.  
+
+<|ref|>text<|/ref|><|det|>[[71, 151, 919, 321]]<|/det|>
+What are the noteworthy results? The authors have done an excellent job of identifying and advancing highly potent inhibitors against their chosen target, cytochrome bc1. There is enough evidence to demonstrate that this is a sound choice of target, especially with the excellent success of bedaquiline, the flagship ATP synthase developed by the same group that has literally changed the paradigm in TB treatment shortening and highlighted the electron transport chain as a highly valuable and sterilizing target, at the same level or better than the rifamycin of prior years. It is therefore sound and admirable to hit this pathway as hard as possible, at multiple points and potentially achieve better cure. The problem seems to be that the interplay between cytochrome bc and cytochrome bd seems complicated and makes it difficult to determine with available data, including the data reported here, if just inhibiting bc, but not bd will be enough to realize the expected benefits. It seems that high expression of bd is able to reverse or nullify the effects of inhibiting bc. This is potentially the reason why there is strain difference between lab adapted versus clinical isolates. Further proof is gleaned from the excellent performance of Q203, the lead asset in this target space, versus mycobacterial species that do not express bd like M. leprae and M. ulcerans. This leaves the potential for a bc inhibitor uncertain against tuberculosis. Having said that, the team has identified a tool compound that together with Q203 will be useful in further investigating this target  
+
+<|ref|>text<|/ref|><|det|>[[72, 332, 920, 439]]<|/det|>
+Will the work be of significance to the field and related fields? While the actual developability of this asset in future TB drug regimens is in doubt until more work is done to fully appreciate contribution against all TB strains, publication of this work will be of great significance to the field and specifically to the grand idea of targeting the electron transport chain of Mtb at multiple nodes to achieve better cure for TB. Furthermore, the compound reported here could also be expanded to the NTM space as is being done for Q203 which is currently being evaluated against leprosy and buruli ulcer, arguably important and neglected diseases as well. But for use in tuberculosis, more work needs to be done to understand the importance of targeting bc1 in a background with high expression of bd. And to determine if the site of infection and location in a lesion affects the usefulness of this asset. And if the mouse model as used today is the best model for this particular asset  
+
+<|ref|>text<|/ref|><|det|>[[72, 449, 921, 568]]<|/det|>
+How does it compare to the established literature? The work reported here is of a high level, similar to other novel drug development programs. Importantly, this was not just a random shot in the dark, instead it was a highly impressive hypothesis driven mission that actually delivered a high- quality compound that is available for further evaluation to prove the hypothesis and potentially deliver a developable asset. It could be argued that in some of the regimen development, drug combination studies in mice, a more factorial approach that builds from 2, to 3 and then perhaps four drugs, with the ability to clearly demonstrate the sterilizing contribution of the asset being evaluated could have been more rigorous than reported here. This leaves this reviewer unable in some cases to confirm that the J bc1 inhibitor contributed anything at all in some of the reported combinations. This could be cleaned up on a re- submission, or in future studies that would attempt to demonstrate the full potential of this asset.  
+
+<|ref|>text<|/ref|><|det|>[[72, 580, 761, 595]]<|/det|>
+If the work is not original, please provide relevant references. The work is original and very well done  
+
+<|ref|>text<|/ref|><|det|>[[72, 606, 921, 648]]<|/det|>
+Does the work support the conclusions and claims, or is additional evidence needed? The work supports some of the claims, but as stated above, the usefulness of developing a bc1 inhibitors and that inhibitor demonstrating a real contribution to treatment shortening against a tuberculosis infection still needs more work.  
+
+<|ref|>text<|/ref|><|det|>[[72, 658, 911, 700]]<|/det|>
+Are there any flaws in the data analysis, interpretation and conclusions? Target validation for bc1 in a highly expressed bd background needs special attention. Contribution of the asset to significant treatment shortening in a developable regimen could be improved  
+
+<|ref|>text<|/ref|><|det|>[[72, 710, 923, 752]]<|/det|>
+Do these prohibit publication or require revision? I recommend publication of this work with a more substantive discussion of the strain difference and usefulness of bc1 inhibitors vs bd expression and a better demonstration of contribution to treatment shortening  
+
+<|ref|>text<|/ref|><|det|>[[72, 762, 723, 778]]<|/det|>
+Is the methodology sound? The methods used are sound except for the identified shortcomings  
+
+<|ref|>text<|/ref|><|det|>[[72, 788, 917, 830]]<|/det|>
+Does the work meet the expected standards in your field? The work is high quality from a highly experienced group working in a target pathway with which they are clearly experts. The regimen development work, especially with this challenging target, relative to bd expression could be improved  
+
+<|ref|>text<|/ref|><|det|>[[72, 840, 635, 855]]<|/det|>
+Is there enough detail provided in the methods for the work to be reproduced? Yes  
+
+<|ref|>text<|/ref|><|det|>[[72, 866, 714, 881]]<|/det|>
+Recommendation: Accept for publication with some revision of the highlighted aspects, above  
+
+<|ref|>text<|/ref|><|det|>[[72, 906, 161, 919]]<|/det|>
+Reviewer #3  
+
+<|ref|>text<|/ref|><|det|>[[72, 931, 238, 945]]<|/det|>
+(Remarks to the Author)
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[73, 48, 208, 60]]<|/det|>
+## General Comments  
+
+<|ref|>text<|/ref|><|det|>[[72, 60, 914, 140]]<|/det|>
+The manuscript presents valuable findings on the role of cytochrome bc1 inhibitors in tuberculosis treatment. However, it currently lacks a clear and well- organized structure, which makes it challenging to follow. To be considered suitable for publication, we strongly recommend rewriting the whole manuscript to include distinct and complete sections—specifically, Introduction, Materials & Methods, Results, Discussion, and Conclusion. The current format hinders readability and clarity and omits critical information necessary to fully assess the study's findings. Additionally, the manuscript would benefit from enhanced experimental justification and a more thorough discussion of the results.  
+
+<|ref|>text<|/ref|><|det|>[[72, 139, 870, 180]]<|/det|>
+The experimental design, including details on mice strains and compound combinations and its rationale, should be consistently presented across all sections (Introduction, Results, Materials & Methods) and thoroughly discussed. Introduction  
+
+<|ref|>text<|/ref|><|det|>[[72, 178, 920, 245]]<|/det|>
+The manuscript provides a well- articulated description of the problem and effectively highlights the potential of the electron transport chain inhibitors in tuberculosis treatment. However, we miss a proper introduction and a clear justification of the two new drug candidates (JNJ- 2901 and JNJ- 4052) that the authors test (according to the figures), and the differences and their potential benefits when compared to the cytochrome bc1 inhibitor Q203 (Telacebec) already being evaluated in clinical trials. This would provide proof of the manuscript's novelty.  
+
+<|ref|>text<|/ref|><|det|>[[72, 243, 912, 283]]<|/det|>
+Hypothesis, aim and objectives are not well presented, and clearly insufficiently described. Moreover, supplementary information and figures should not be included in the Introduction but rather integrated and discussed in the results section. Results  
+
+<|ref|>text<|/ref|><|det|>[[72, 281, 916, 323]]<|/det|>
+The results demonstrate promising data on JNJ- 2901's efficacy in reducing bacterial burden and preventing relapse in mice models. The findings suggest a potential shortening of TB treatment regimens, particularly with BPaCJ, which showed the highest efficacy and lowest relapse rates compared to BPaL, BPaM, BPaC, and BPaJ.  
+
+<|ref|>text<|/ref|><|det|>[[72, 322, 683, 336]]<|/det|>
+However, we have major concerns regarding the description and structure of this section:  
+
+<|ref|>text<|/ref|><|det|>[[72, 335, 920, 440]]<|/det|>
+- Global experimental plan should first be presented to help understand what the authors did, and then each set of results should be clearly introduced with a summary of the main conclusion before presenting the data of each experimental study. 
+- The experimental design (mice strain, infection mode, treatment duration, and rationale for each study) and must be clearly stated at the beginning of each results subsection. We here wanted to point out that the naming convention of studies (e.g., Study A, B, C, etc.) does not provide enough information to the reader. We suggest their replacement throughout the manuscript with descriptive terms reflecting the experimental model. For example, instead of "In an additional study (Study C)," we suggest "We additionally investigated drug combination efficacy and relapse in intravenously infected mice." 
+- Why a second compound – JNJ-4052- is tested in Study F? Why to test another candidate?  
+
+<|ref|>text<|/ref|><|det|>[[72, 439, 152, 451]]<|/det|>
+Discussion  
+
+<|ref|>text<|/ref|><|det|>[[72, 451, 918, 491]]<|/det|>
+Although a very small discussion comments are integrated into the manuscript (e.g., L136- 143, L156- 161, L171- 177), a dedicated Discussion section is necessary. We suggest moving the relevant lines into a standalone Discussion section and expanding on several key topics:  
+
+<|ref|>text<|/ref|><|det|>[[72, 490, 916, 544]]<|/det|>
+- Comparison of JNJ-2901 and JNJ-4052 with Telacebec (Q203): The manuscript should provide a comparative analysis of the new candidates with existing inhibitors in terms of efficacy, safety, and pharmacokinetics as well as treatment efficacy and relapsing rates. What is the novelty? Why these compounds are better? Why they test 2 different compounds? And if everything is done with JNJ-2901, why to include JNJ-4052 in one of the studies?  
+
+<|ref|>text<|/ref|><|det|>[[72, 543, 916, 583]]<|/det|>
+- Differences in mice models and infection routes: The study utilizes different infection models, and these differences should be considered to discuss the results. In addition, the authors should also discuss the gender bias in the experiments (only female mice are used) and how this could impact clinical translation.  
+
+<|ref|>text<|/ref|><|det|>[[72, 582, 916, 622]]<|/det|>
+- Comparison between clinical isolates and wild-type (WT) strains: Authors should also discuss the observed differences in drug efficacy between clinical isolates and WT strains. The reasons behind these differences need to be explained and appropriately discussed.  
+
+<|ref|>text<|/ref|><|det|>[[72, 620, 860, 648]]<|/det|>
+- Experimental limitations: The mention of limitations in L119 is too vague. The authors should explicitly outline the limitations of their experimental design.  
+
+<|ref|>text<|/ref|><|det|>[[72, 647, 916, 700]]<|/det|>
+- The potential of cytochrome bc1 inhibitors: The claim that these inhibitors could significantly impact TB treatment is overstated for the level of preclinical data presented. Further discussion on the limitations of the preclinical findings hereby presented and next steps for clinical translation is needed, including challenges, barriers and further work to be done before undergoing clinical evaluation.  
+
+<|ref|>sub_title<|/ref|><|det|>[[72, 712, 231, 724]]<|/det|>
+## Materials and Methods  
+
+<|ref|>text<|/ref|><|det|>[[72, 724, 916, 750]]<|/det|>
+The Materials & Methods section is well- detailed and allows for experimental reproducibility. However, there we have some comments:  
+
+<|ref|>text<|/ref|><|det|>[[72, 750, 920, 790]]<|/det|>
+- As in the Results section, avoid referring only to different experiments as Study A, B, etc. Instead, add descriptive headings such as "Intranasal Mice Challenge", "Intravenous Mice Challenge", or "High-Dose Aerosol Challenge."  
+
+<|ref|>text<|/ref|><|det|>[[72, 789, 800, 803]]<|/det|>
+- Each experimental model should consistently include the following information: Mouse strain, sex, and age; inoculum  
+
+<|ref|>text<|/ref|><|det|>[[72, 802, 655, 815]]<|/det|>
+preparation, infection route, and dose; treatment duration and euthanasia time points.  
+
+<|ref|>text<|/ref|><|det|>[[72, 814, 551, 827]]<|/det|>
+- Authors should provide in the ethical statement the end-point criteria.  
+
+<|ref|>text<|/ref|><|det|>[[72, 826, 920, 852]]<|/det|>
+- Supplementary tables and results should not be included in the Methods section but should be properly cited in the Results section.  
+
+<|ref|>text<|/ref|><|det|>[[72, 850, 910, 910]]<|/det|>
+- There are several experimental methods referenced in the supplementary material but not appearing in the main text (e.g. PK and tolerability in mouse, ADME assays, HepG2 cytotoxicity assay, mitochondrial toxicity assay Glu/Gal, Ames II Mutagenicity assay, mutant isolation and WGS). These should be briefly mentioned in the main text, such as: "Both JNJ-2901 and JNJ-4052 were evaluated for ADME and toxicity. All parameters supported further use of these compounds in subsequent experiments (Supplementary Table S2)".  
+
+<|ref|>text<|/ref|><|det|>[[72, 909, 530, 922]]<|/det|>
+Additionally, we suggest to address the following minor comments:  
+
+<|ref|>text<|/ref|><|det|>[[72, 922, 560, 948]]<|/det|>
+- L62: we would suggest classifying tuberculosis as a global pandemic. 
+- L79: MoA abbreviation is not needed.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[70, 46, 914, 87]]<|/det|>
+- L87: the word "validated" to describe the inhibitor compound JNJ-2901 is confusing as the presented manuscript seems to be the validation of the compound as an effective treatment adjuvant. If this is not the case, please provide the reference where the compound is validated for its inhibition of the cytochrome bc1.  
+
+<|ref|>text<|/ref|><|det|>[[70, 87, 910, 113]]<|/det|>
+- L91-96: we suggest to relocate this in the line 76, as the TB-PRACTECAL trial results are a good justification of linezolid-associated adverse effects when treating MDR-TB.  
+
+<|ref|>text<|/ref|><|det|>[[73, 113, 465, 126]]<|/det|>
+- L116-122: please contextualize the results from this part.  
+
+<|ref|>text<|/ref|><|det|>[[73, 126, 576, 139]]<|/det|>
+- L127: move the figure citation to line 133. After the results are explained.  
+
+<|ref|>text<|/ref|><|det|>[[70, 140, 905, 166]]<|/det|>
+- L146: please provide the justification of why using Telacobec in this study and not JNJ-2901 and provide rationale about the drug regimen used for this experimental design.  
+
+<|ref|>text<|/ref|><|det|>[[70, 166, 900, 205]]<|/det|>
+- L164: as far as we understood, JNJ-2901 and Telacobec were also assessed for its bactericidal activity against clinical isolates in comparison to the H37Rv strain (Figure 2 B,C,D and E). Please provide a proper description of the compound used for this experiment and further discussion of this results in the main text.  
+
+<|ref|>text<|/ref|><|det|>[[70, 204, 884, 230]]<|/det|>
+- Figure 1: If the mice strain and experimental design are the same, consider re-structure Figure 1B and 1C in a unique figure.  
+
+<|ref|>text<|/ref|><|det|>[[70, 230, 918, 270]]<|/det|>
+- Figure 2: Please, consider a figure rearrangement. We suggest dividing Figure 2 in two. Section A on one side (resulting in Figure 2) and sections B, C, D, E and F on the other (resulting in Figure 3 A, B, C, D and E). Since the results represented refer to different experimental questions and require a results section for each one.  
+
+<|ref|>sub_title<|/ref|><|det|>[[72, 283, 235, 295]]<|/det|>
+## Final Recommendation  
+
+<|ref|>text<|/ref|><|det|>[[72, 296, 920, 348]]<|/det|>
+This study presents valuable preclinical data on cytochrome bc1 inhibitors in TB treatment. However, we have significant concerns regarding the manuscript as it currently stands. It lacks a proper structure, clarity, and experimental justification, and the discussion is entirely insufficient. The manuscript needs to be totally rewritten to address these issues, ensuring that it achieves the required readability, scientific rigor, and overall impact to be considered suitable for publication.  
+
+<|ref|>sub_title<|/ref|><|det|>[[72, 373, 161, 386]]<|/det|>
+## Reviewer #4  
+
+<|ref|>text<|/ref|><|det|>[[72, 399, 238, 411]]<|/det|>
+(Remarks to the Author)  
+
+<|ref|>text<|/ref|><|det|>[[72, 412, 864, 451]]<|/det|>
+I co- reviewed this manuscript with one of the reviewers who provided the listed reports. This is part of the Nature Communications initiative to facilitate training in peer review and to provide appropriate recognition for Early Career Researchers who co- review manuscripts.  
+
+<|ref|>text<|/ref|><|det|>[[72, 464, 144, 476]]<|/det|>
+Version 1:  
+
+<|ref|>text<|/ref|><|det|>[[72, 490, 219, 503]]<|/det|>
+Reviewer comments:  
+
+<|ref|>text<|/ref|><|det|>[[72, 515, 160, 528]]<|/det|>
+Reviewer #1  
+
+<|ref|>text<|/ref|><|det|>[[72, 542, 238, 554]]<|/det|>
+(Remarks to the Author)  
+
+<|ref|>text<|/ref|><|det|>[[72, 555, 923, 595]]<|/det|>
+The revised manuscript of Clara Aguilar Pérez and colleagues has addressed all of my previous concerns and I am happy to see some of the recommendation on the revised discussion section as well as extended new data. The study represents an important progress on inhibition of respiration machinery in Mtb and has potential for the development of future therapeutics.  
+
+<|ref|>text<|/ref|><|det|>[[72, 607, 160, 620]]<|/det|>
+Reviewer #3  
+
+<|ref|>text<|/ref|><|det|>[[72, 634, 238, 646]]<|/det|>
+(Remarks to the Author)  
+
+<|ref|>text<|/ref|><|det|>[[72, 647, 915, 712]]<|/det|>
+We would like to thank the authors for thoughtfully addressing our major suggestions regarding the manuscript. Our feedback primarily focused on improving its structure, clarity, and the discussion surrounding the experimental procedures and results. In particular, the authors have now effectively justified the differences observed when treating laboratory strains with different cytochrome bc1 inhibitors versus clinical strains. With these revisions, the manuscript now clearly highlights the significance of this investigation in advancing our understanding of tuberculosis treatment strategies.  
+
+<|ref|>text<|/ref|><|det|>[[72, 724, 920, 802]]<|/det|>
+This work underscores the potential of bc1 inhibitors as viable alternatives to fluoroquinolones and linezolid, which are often associated with adverse effects in current regimens for drug- resistant tuberculosis. Moreover, it contributes to the possibility of treatment shortening for drug- sensitive tuberculosis, with added potential against clinical strains. The experiments were performed using different, well- justified preclinical mouse models, which strengthens the translational relevance of the findings and supports further investigation of this promising asset. Based on these improvements and the scientific merit of the study, we believe the manuscript is now better suited for publication.  
+
+<|ref|>text<|/ref|><|det|>[[72, 826, 602, 840]]<|/det|>
+However, we would like the authors to address the following remaining issue:  
+
+<|ref|>text<|/ref|><|det|>[[72, 853, 920, 945]]<|/det|>
+Please add a paragraph on limitations to the discussion section. For example, the sample size of mice used in the experiments could be acknowledged as a limitation (even though we recognize that this is already mentioned on page 7). Additionally, regarding the authors' response that "no differences between sexes have been reported in relation to drug efficacy in TB mouse models," we would like to stress that, even if ethical requirements do not mandate the inclusion of both biological sexes, doing so is strongly advisable. There is robust evidence from both clinical and preclinical studies that biological sex influences TB epidemiology and treatment outcomes (for example, see: Idris, R., Infection 2025, https://doi.org/10.1007/s15010-024-02424-5; Dutta NK, Front Immunol, 2020, doi: 10.3389/fimmu.2020.01465; Tannenbaum,
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[72, 46, 921, 166]]<|/det|>
+C., Nature, 2019, https://doi.org/10.1038/s41586-019-1657-6). These differences are driven by a combination of hormonal, genetic, and immunological factors, as well as potential differences in treatment adherence and broader social determinants of health. While we acknowledge that using female mice can be more practical in terms of logistics and animal management- since males often display more aggressive behavior that can complicate experiments- there is now ample evidence showing that men and women often have different disease trajectories (not only for TB!) and responses to treatment. One persistent issue in drug development is that preclinical and clinical studies frequently fail to reflect the biological diversity found in real- world populations. Including both sexes in preclinical experiments wherever feasible would help improve the generalizability of the findings and ensure that results more accurately inform treatment strategies for all patients.  
+
+<|ref|>text<|/ref|><|det|>[[70, 177, 905, 206]]<|/det|>
+In summary, we encourage the authors to carefully consider any additional limitations of their study and clearly state them, as this will add valuable context and transparency. Once this is done, the manuscript can be endorsed for publication.  
+
+<|ref|>sub_title<|/ref|><|det|>[[73, 216, 162, 230]]<|/det|>
+## Reviewer #4  
+
+<|ref|>text<|/ref|><|det|>[[73, 243, 863, 296]]<|/det|>
+(Remarks to the Author) I co- reviewed this manuscript with one of the reviewers who provided the listed reports. This is part of the Nature Communications initiative to facilitate training in peer review and to provide appropriate recognition for Early Career Researchers who co- review manuscripts.  
+
+<|ref|>text<|/ref|><|det|>[[72, 650, 916, 702]]<|/det|>
+Open Access This Peer Review File is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.  
+
+<|ref|>text<|/ref|><|det|>[[72, 702, 915, 768]]<|/det|>
+In cases where reviewers are anonymous, credit should be given to 'Anonymous Referee' and the source. The images or other third party material in this Peer Review File are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.  
+
+<|ref|>text<|/ref|><|det|>[[73, 767, 618, 780]]<|/det|>
+To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[105, 81, 294, 98]]<|/det|>
+## Reviewers' comments  
+
+<|ref|>text<|/ref|><|det|>[[102, 107, 894, 143]]<|/det|>
+We thank the reviewers for providing useful comments and suggestions that have greatly improved this manuscript. We have addressed each comment below (red).  
+
+<|ref|>sub_title<|/ref|><|det|>[[105, 180, 425, 198]]<|/det|>
+## Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[104, 206, 894, 311]]<|/det|>
+The manuscript by Perez and colleagues present data on the effects of cytochrome bc1 inhibitors and their potential efficacy in drug regimens against multidrug resistant TB as well as its role in reducing the duration of chemotherapy in drug susceptible TB. The work is important and timely considered the burden of drug- resistant TB and the prolonged treatment protocols for both drug resistant and susceptible cases of TB. The experimental approach is solid and data presented are robust to support their hypotheses.  
+
+<|ref|>text<|/ref|><|det|>[[104, 319, 894, 406]]<|/det|>
+One of the main interesting points in the manuscript is the fact that cytochrome bc1 inhibitors are more potent against clinical isolates than the lab strains. A further investigation into this would have increased the impact of the work as well as our understanding of the mechanisms of action of these inhibitors. Especially as these are claimed to be validated inhibitors, it would also be nice to see more of their characterisation and target engagement in vitro.  
+
+<|ref|>text<|/ref|><|det|>[[104, 415, 894, 467]]<|/det|>
+A second paper has recently been published which describes the characterisation, validation and target engagement of the JNJ- 2901 inhibitor (PMID:40191462). We have provided validation of JNJ- 4052, which comes from the same chemical series as JNJ- 2901, in Supplementary Tables S1- 3.  
+
+<|ref|>text<|/ref|><|det|>[[104, 478, 894, 530]]<|/det|>
+In addition, would this variability on strain efficacy mean that there are potential other targets being inhibited by the same compounds or is it the secondary effects of inhibiting the cytochrome bc1 and the respiratory chain of Mtb?  
+
+<|ref|>text<|/ref|><|det|>[[104, 539, 894, 643]]<|/det|>
+We do not believe that there are any genuine secondary targets for these compounds. Our resistance data demonstrates that a single nucleotide polymorphism in qcrB can provide high level resistance (for example T313A leads to \(>2000\) - fold resistance to JNJ- 4052; Supplementary Table S1). As discussed below, we believe the differences are most likely due to the increased expression of cytochrome bd, an alternative oxidase that can compensate for cytochrome bc1 inhibition. This is now outlined in the Discussion section (pg 9- 10).  
+
+<|ref|>text<|/ref|><|det|>[[104, 652, 894, 790]]<|/det|>
+For the present study, authors can expand the discussion regarding the efficacy of their compound in the background of bd and how this could affect the outcome. More importantly, the authors can use these compounds to further interrogate this hypothesis. In addition, having another look at their data, I could see that if you calculate the MBC/MIC ratio for the reference lab strain (H37Rv) and the cytochrome deleted mutant (H37Rv_CytBd- KO) on supplementary table S2 one can see that their inhibitor is bacteriostatic on H37Rv whereas it is bactericidal on cytochrome bd deleted mutant (H37Rv_CytBd- KO). That further adds to the argument that the expression levels of cytochrome bd is an important factor in their compound efficacy.  
+
+<|ref|>text<|/ref|><|det|>[[104, 800, 894, 835]]<|/det|>
+We agree that this is an important point and have now provided discussion of the difference in bactericidality between H37Rv and H37Rv_CytBd- KO strains in the Discussion section (pg 9- 10).  
+
+<|ref|>text<|/ref|><|det|>[[104, 844, 894, 914]]<|/det|>
+Furthermore, the authors did not include if these compounds are active against MDR and XDR Mtb strains. Taking into account that the main use of these inhibitors would be towards a drug regimen for treating MDR- TB, it would be nice to see that the compounds maintain their activity against a panel of multiple drug- resistant strains including MDR and XDR clinical isolates.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[102, 80, 894, 115]]<|/det|>
+Activity of JNJ- 2901 against a panel of 18 MDR- TB strains has now been published showing that the inhibitor is more active compared to lab- adapted strains (PMID:40191462).  
+
+<|ref|>text<|/ref|><|det|>[[104, 124, 894, 177]]<|/det|>
+Obviously a series of experiments where one looks at the levels of expression of these genes under treatment or not in a series of lab and clinical strains of Mtb would shed light on this question. But I don't think these experiments need to be part of this work.  
+
+<|ref|>text<|/ref|><|det|>[[104, 186, 894, 221]]<|/det|>
+Lastly, in the Figure S3, I cannot see the control line as well as the figure legend need a bit more explanation.  
+
+<|ref|>text<|/ref|><|det|>[[104, 231, 894, 266]]<|/det|>
+We have updated the figure legend to indicate that moxifloxacin is hidden by BDQ (Supplementary Figure S3). We have added more detail of the experimental conditions in the figure legend.  
+
+<|ref|>sub_title<|/ref|><|det|>[[105, 310, 425, 328]]<|/det|>
+## Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[103, 337, 894, 562]]<|/det|>
+The authors have done an excellent job of identifying cytochrome bc1 as an important component of the Mtb electron transport chain and a potentially sterilizing target in for TB regimen development. Using a relapsing mouse model, they have attempted to demonstrate that cytochrome bc1 inhibitors could be effective partner drugs in TB regimen development that could enhance sterilization. They report on several novel regimens as examples of this role for bc1 inhibitors in regimens for both multidrug- resistant TB (MDR- TB) and drug- sensitive TB (DS- TB), where cytochrome bc1 inhibitors could contribute to sterilization and treatment shortening. importantly, they have shown that clinical isolates exhibit heightened susceptibility to cytochrome bc1 inhibitors compared to laboratory- adapted strains, further supporting their potential usefulness in a clinical setting even as their contribution may be low or undetectable against a lab adapted strain. They take these findings as confirmation that cytochrome bc1 inhibitors have significant potential to improve TB treatment outcomes and highlight the need for further studies to evaluate their clinical contribution to novel treatment regimens.  
+
+<|ref|>text<|/ref|><|det|>[[103, 571, 894, 866]]<|/det|>
+What are the noteworthy results? The authors have done an excellent job of identifying and advancing highly potent inhibitors against their chosen target, cytochrome bc1. There is enough evidence to demonstrate that this is a sound choice of target, especially with the excellent success of bedaquiline, the flagship ATP synthase developed by the same group that has literally changed the paradigm in TB treatment shortening and highlighted the electron transport chain as a highly valuable and sterilizing target, at the same level or better than the rifamycins of prior years. It is therefore sound and admirable to hit this pathway as hard as possible, at multiple points and potentially achieve better cure. The problem seems to be that the interplay between cytochrome bc and cytochrome bd seems complicated and makes it difficult to determine with available data, including the data reported here, if just inhibiting bc, but not bd will be enough to realize the expected benefits. It seems that high expression of bd is able to reverse or nullify the effects of inhibiting bc. This is potentially the reason why there is strain difference between lab adapted versus clinical isolates. Further proof is gleaned from the excellent performance of Q203, the lead asset in this target space, versus mycobacterial species that do not express bd like M. leprae and M. ulcerans. This leaves the potential for a bc inhibitor uncertain against tuberculosis. Having said that, the team has identified a tool compound that together with Q203 will be useful in further investigating this target  
+
+<|ref|>text<|/ref|><|det|>[[103, 875, 894, 910]]<|/det|>
+Will the work be of significance to the field and related fields? While the actual developability of this asset in future TB drug regimens is in doubt until more work is done to fully appreciate contribution
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[104, 80, 893, 219]]<|/det|>
+against all TB strains, publication of this work will be of great significance to the field and specifically to the grand idea of targeting the electron transport chain of Mtb at multiple nodes to achieve better cure for TB. Furthermore, the compound reported here could also be expanded to the NTM space as is being done for Q203 which is currently being evaluated against leprosy and buruli ulcer, arguably important and neglected diseases as well. But for use in tuberculosis, more work needs to be done to understand the importance of targeting bc1 in a background with high expression of bd. And to determine if the site of infection and location in a lesion affects the usefulness of this asset. And if the mouse model as used today is the best model for this particular asset  
+
+<|ref|>text<|/ref|><|det|>[[104, 228, 894, 418]]<|/det|>
+How does it compare to the established literature? The work reported here is of a high level, similar to other novel drug development programs. Importantly, this was not just a random shot in the dark, instead it was a highly impressive hypothesis driven mission that actually delivered a high- quality compound that is available for further evaluation to prove the hypothesis and potentially deliver a developable asset. It could be argued that in some of the regimen development, drug combination studies in mice, a more factorial approach that builds from 2, to 3 and then perhaps four drugs, with the ability to clearly demonstrate the sterilizing contribution of the asset being evaluated could have been more rigorous than reported here. This leaves this reviewer unable in some cases to confirm that the J bc1 inhibitor contributed anything at all in some of the reported combinations. This could be cleaned up on a re- submission, or in future studies that would attempt to demonstrate the full potential of this asset.  
+
+<|ref|>text<|/ref|><|det|>[[104, 428, 894, 531]]<|/det|>
+A more stepwise factorial approach (building from two- to three- and then four- drug combinations) could have further clarified the sterilising contribution of a cytochrome \(bc_{1}\) inhibitor. However, the aim of these studies was not to develop an entirely new regimen but rather to replace moxifloxacin and linezolid with a potentially superior compound while maintaining the existing backbone. Therefore, we prioritised evaluating the \(bc_{1}\) inhibitor in combination with a clinically relevant regimen to reflect a realistic therapeutic setting.  
+
+<|ref|>text<|/ref|><|det|>[[105, 541, 893, 592]]<|/det|>
+Additionally, we were mindful of the ethical considerations regarding animal use. Expanding the study design to include a larger number of combinations/conditions would have significantly increased the number of animals required.  
+
+<|ref|>text<|/ref|><|det|>[[105, 602, 893, 688]]<|/det|>
+In several of our studies, the addition of a cytochrome \(bc_{1}\) inhibitor led to a statistical improvement in relapse rates (e.g. Figure 2; CZ vs CZT). Nevertheless, we agree that future investigations could employ a more granular factorial design to better define the specific contribution of the \(bc_{1}\) inhibitor. These follow- up studies would help further demonstrate the full potential of cytochrome \(bc_{1}\) inhibitors and strengthen its case for clinical development.  
+
+<|ref|>text<|/ref|><|det|>[[105, 699, 892, 716]]<|/det|>
+If the work is not original, please provide relevant references. The work is original and very well done  
+
+<|ref|>text<|/ref|><|det|>[[105, 726, 893, 794]]<|/det|>
+Does the work support the conclusions and claims, or is additional evidence needed? The work supports some of the claims, but as stated above, the usefulness of developing a bc1 inhibitors and that inhibitor demonstrating a real contribution to treatment shortening against a tuberculosis infection still needs more work.  
+
+<|ref|>text<|/ref|><|det|>[[105, 806, 893, 908]]<|/det|>
+While we acknowledge that additional studies are needed to fully establish the potential of cytochrome \(bc_{1}\) inhibitors to shorten treatment duration, particularly in the context of multidrug- resistant tuberculosis (MDR- TB), the data presented here clearly demonstrate that the inclusion of a cytochrome \(bc_{1}\) inhibitor reduces the treatment time required. For example, the addition of telacebec to CZ reduces the treatment by at least 2 months (Figure 2). This is the starting point for further work to investigate this in more detail.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[102, 97, 895, 132]]<|/det|>
+Are there any flaws in the data analysis, interpretation and conclusions? Target validation for bc1 in a highly expressed bd background needs special attention.  
+
+<|ref|>text<|/ref|><|det|>[[102, 141, 895, 177]]<|/det|>
+We have recently published the validation of JNJ- 2901 (PMID: 40191462) and have expanded our discussion on bd expression (pg 9- 10).  
+
+<|ref|>text<|/ref|><|det|>[[102, 186, 895, 221]]<|/det|>
+Contribution of the asset to significant treatment shortening in a developable regimen could be improved  
+
+<|ref|>text<|/ref|><|det|>[[104, 231, 895, 334]]<|/det|>
+As discussed above, we view this work as an important starting point for investigating the potential contribution of cytochrome \(bc_{1}\) inhibitors to future treatment- shortening regimens. We agree that further work is required beyond the current study. However, one challenge in addressing this more definitively is the lack of consensus on what constitutes a 'developable' regimen, as different organisations often have varying perspectives on which drugs should be prioritised; clofazimine being a notable example here.  
+
+<|ref|>text<|/ref|><|det|>[[104, 344, 895, 397]]<|/det|>
+Do these prohibit publication or require revision? I recommend publication of this work with a more substantive discussion of the strain difference and usefulness of bc1 inhibitors vs bd expression and a better demonstration of contribution to treatment shortening  
+
+<|ref|>text<|/ref|><|det|>[[104, 407, 850, 441]]<|/det|>
+We have now re- formatted the manuscript and expanded the discussion on the clinical isolate differences we observed.  
+
+<|ref|>text<|/ref|><|det|>[[102, 468, 864, 486]]<|/det|>
+Is the methodology sound? The methods used are sound except for the identified shortcomings  
+
+<|ref|>text<|/ref|><|det|>[[104, 495, 870, 565]]<|/det|>
+Does the work meet the expected standards in your field? The work is high quality from a highly experienced group working in a target pathway with which they are clearly experts. The regimen development work, especially with this challenging target, relative to bd expression could be improved  
+
+<|ref|>text<|/ref|><|det|>[[104, 565, 850, 600]]<|/det|>
+Is there enough detail provided in the methods for the work to be reproduced? Yes Recommendation: Accept for publication with some revision of the highlighted aspects, above  
+
+<|ref|>sub_title<|/ref|><|det|>[[104, 643, 487, 660]]<|/det|>
+## Reviewer #3 and #4 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[105, 661, 263, 676]]<|/det|>
+General Comments  
+
+<|ref|>text<|/ref|><|det|>[[104, 678, 895, 816]]<|/det|>
+The manuscript presents valuable findings on the role of cytochrome bc1 inhibitors in tuberculosis treatment. However, it currently lacks a clear and well- organized structure, which makes it challenging to follow. To be considered suitable for publication, we strongly recommend rewriting the whole manuscript to include distinct and complete sections—specifically, Introduction, Materials & Methods, Results, Discussion, and Conclusion. The current format hinders readability and clarity and omits critical information necessary to fully assess the study's findings. Additionally, the manuscript would benefit from enhanced experimental justification and a more thorough discussion of the results.  
+
+<|ref|>text<|/ref|><|det|>[[104, 825, 895, 860]]<|/det|>
+We have now re- formatted the manuscript for Nature Communications as suggested and expanded on the Discussion section.  
+
+<|ref|>text<|/ref|><|det|>[[102, 887, 860, 904]]<|/det|>
+The experimental design, including details on mice strains and compound combinations and its
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[104, 80, 875, 115]]<|/det|>
+rationale, should be consistently presented across all sections (Introduction, Results, Materials & Methods) and thoroughly discussed.  
+
+<|ref|>sub_title<|/ref|><|det|>[[104, 143, 201, 159]]<|/det|>
+## Introduction  
+
+<|ref|>text<|/ref|><|det|>[[104, 160, 888, 264]]<|/det|>
+The manuscript provides a well- articulated description of the problem and effectively highlights the potential of the electron transport chain inhibitors in tuberculosis treatment. However, we miss a proper introduction and a clear justification of the two new drug candidates (JNJ- 2901 and JNJ- 4052) that the authors test (according to the figures), and the differences and their potential benefits when compared to the cytochrome bc1 inhibitor Q203 (Telacebec) already being evaluated in clinical trials. This would provide proof of the manuscript's novelty.  
+
+<|ref|>text<|/ref|><|det|>[[104, 273, 895, 378]]<|/det|>
+The primary focus of this manuscript is to understand the contribution of a cytochrome bc1- targeting inhibitor to TB treatment regimens, rather than specifically focusing on JNJ- 2091 or suggesting it is a superior molecule. We consider all the compounds discussed in this work as tool compounds used to increase our understanding of how this mode of action could complement existing drug regimens. In parallel, we have published a separate validation of JNJ- 2901 (PMID:40191462) and have added additional justification here (pg 5- 6).  
+
+<|ref|>text<|/ref|><|det|>[[104, 386, 822, 404]]<|/det|>
+Hypothesis, aim and objectives are not well presented, and clearly insufficiently described.  
+
+<|ref|>text<|/ref|><|det|>[[105, 413, 715, 431]]<|/det|>
+We have clarified and expanded the objectives of each section as requested.  
+
+<|ref|>text<|/ref|><|det|>[[104, 440, 864, 476]]<|/det|>
+Moreover, supplementary information and figures should not be included in the Introduction but rather integrated and discussed in the results section.  
+
+<|ref|>text<|/ref|><|det|>[[104, 485, 894, 520]]<|/det|>
+Reference to the figures and supplementary information has been removed from the introduction following re- formatting.  
+
+<|ref|>sub_title<|/ref|><|det|>[[104, 549, 167, 564]]<|/det|>
+## Results  
+
+<|ref|>text<|/ref|><|det|>[[104, 565, 872, 633]]<|/det|>
+The results demonstrate promising data on JNJ- 2901's efficacy in reducing bacterial burden and preventing relapse in mice models. The findings suggest a potential shortening of TB treatment regimens, particularly with BPaCJ, which showed the highest efficacy and lowest relapse rates compared to BPaL, BPaM, BPaC, and BPaJ.  
+
+<|ref|>text<|/ref|><|det|>[[105, 633, 812, 650]]<|/det|>
+However, we have major concerns regarding the description and structure of this section:  
+
+<|ref|>text<|/ref|><|det|>[[104, 651, 892, 703]]<|/det|>
+- Global experimental plan should first be presented to help understand what the authors did, and then each set of results should be clearly introduced with a summary of the main conclusion before presenting the data of each experimental study.  
+
+<|ref|>text<|/ref|><|det|>[[104, 712, 893, 747]]<|/det|>
+We agree, and have now divided this section more clearly with Figures 1b- c now discussed sequentially. An overview of the first set of experiments is provided in Figure 1a.  
+
+<|ref|>text<|/ref|><|det|>[[104, 756, 895, 877]]<|/det|>
+- The experimental design (mice strain, infection mode, treatment duration, and rationale for each study) and must be clearly stated at the beginning of each results subsection. We here wanted to point out that the naming convention of studies (e.g., Study A, B, C, etc.) does not provide enough information to the reader. We suggest their replacement throughout the manuscript with descriptive terms reflecting the experimental model. For example, instead of “In an additional study (Study C),” we suggest “We additionally investigated drug combination efficacy and relapse in intravenously infected mice.”
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[104, 80, 894, 150]]<|/det|>
+We respectfully disagree, as the experimental design is indicated in the Materials and Methods section. We do not believe it is necessary to repeat this information in the Results section or figure legends. We have named them A, B, C etc. because the studies were performed by different institutions rather than because they are different type of studies.  
+
+<|ref|>text<|/ref|><|det|>[[104, 159, 840, 177]]<|/det|>
+- Why a second compound – JNJ-4052- is tested in Study F? Why to test another candidate?  
+
+<|ref|>text<|/ref|><|det|>[[104, 186, 894, 308]]<|/det|>
+JNJ- 2901 and JNJ- 4052 belong to the same chemical series and exhibit very similar activity and safety profiles. Our cytochrome \(bc_{1}\) drug discovery programme has been ongoing for 7 years, during which time we have progressed a number of lead compounds including JNJ- 4052 and JNJ- 2901. The in vivo clinical isolate study (Study F) was performed using the earlier lead, JNJ- 4052. While we acknowledge that repeating this experiment with JNJ- 2901 would have been ideal, given the comparable profiles of both compounds, we do not anticipate a change in the outcome. Therefore, we believe repeating this study would not represent a justifiable use of mice.  
+
+<|ref|>sub_title<|/ref|><|det|>[[104, 318, 194, 334]]<|/det|>
+## Discussion  
+
+<|ref|>text<|/ref|><|det|>[[104, 336, 888, 400]]<|/det|>
+Although a very small discussion comments are integrated into the manuscript (e.g., L136- 143, L156- 161, L171- 177), a dedicated Discussion section is necessary. We suggest moving the relevant lines into a standalone Discussion section and expanding on several key topics:  
+
+<|ref|>text<|/ref|><|det|>[[104, 389, 888, 473]]<|/det|>
+- Comparison of JNJ-2901 and JNJ-4052 with Telacebec (Q203): The manuscript should provide a comparative analysis of the new candidates with existing inhibitors in terms of efficacy, safety, and pharmacokinetics as well as treatment efficacy and relapsing rates. What is the novelty? Why these compounds are better? Why they test 2 different compounds? And if everything is done with JNJ-2901, why to include JNJ-4052 in one of the studies?  
+
+<|ref|>text<|/ref|><|det|>[[104, 483, 894, 552]]<|/det|>
+The aim of the project was to investigate the role of cytochrome \(bc_{1}\) inhibitors in future treatment regimens rather than suggest that JNJ- 2901 is superior to telacebec. We selected three inhibitors with related chemical structure, mode of binding and efficacy to achieve these aims; we have made this clearer in the text (pg 5- 6).  
+
+<|ref|>text<|/ref|><|det|>[[104, 561, 894, 630]]<|/det|>
+- Differences in mice models and infection routes: The study utilizes different infection models, and these differences should be considered to discuss the results. In addition, the authors should also discuss the gender bias in the experiments (only female mice are used) and how this could impact clinical translation.  
+
+<|ref|>text<|/ref|><|det|>[[104, 641, 894, 744]]<|/det|>
+In this study, two different infection models were used: nasal (via aerosol or intranasal instillation) and intravenous. As noted in the text, we acknowledge that the intravenous model exhibits a higher severity of the disease since this infection route confers a systemic infection, whereas the nasal route mainly leads to a pulmonary infection. The intention of this study was not to characterise the treatment response at the pathology level, but rather to have a broad knowledge of the activity of cytochrome \(bc_{1}\) - containing regimens in multiple potential scenarios.  
+
+<|ref|>text<|/ref|><|det|>[[104, 753, 894, 857]]<|/det|>
+Regarding the use of only female mice, we acknowledge recommendations for including both sexes in animal studies. However, in line with our institutional policies focused on the 3Rs, we used only female mice to minimise stress and harm. Relapse studies are very long and involve repeated daily handling and dosing, which can lead to a higher risk of aggres behaviour in males. Clinical translation would take place when human exposures are known in a Phase I trial. Additionally, no differences between sex have been reported in relation to drug efficacy in TB mouse models.  
+
+<|ref|>text<|/ref|><|det|>[[104, 885, 888, 902]]<|/det|>
+- Comparison between clinical isolates and wild-type (WT) strains: Authors should also discuss the
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[102, 80, 890, 115]]<|/det|>
+observed differences in drug efficacy between clinical isolates and WT strains. The reasons behind these differences need to be explained and appropriately discussed.  
+
+<|ref|>text<|/ref|><|det|>[[102, 124, 790, 142]]<|/det|>
+We have now added discussion on the comparison of WT vs. clinical isolate (pg 9- 10).  
+
+<|ref|>text<|/ref|><|det|>[[102, 152, 856, 187]]<|/det|>
+- Experimental limitations: The mention of limitations in L119 is too vague. The authors should explicitly outline the limitations of their experimental design.  
+
+<|ref|>text<|/ref|><|det|>[[102, 196, 890, 232]]<|/det|>
+The main limitation of this study was the number of animals used did not provide enough statistical power to differentiate the different conditions. We have mentioned this in the text (pg 7).  
+
+<|ref|>text<|/ref|><|det|>[[102, 241, 888, 328]]<|/det|>
+- The potential of cytochrome bc1 inhibitors: The claim that these inhibitors could significantly impact TB treatment is overstated for the level of preclinical data presented. Further discussion on the limitations of the preclinical findings hereby presented and next steps for clinical translation is needed, including challenges, barriers and further work to be done before undergoing clinical evaluation.  
+
+<|ref|>text<|/ref|><|det|>[[102, 337, 894, 407]]<|/det|>
+As discussed above, we wish to make it clear that JNJ- 2901 is a tool compound that we have used to demonstrate how, inhibitors of cytochrome bc1, a novel mode of action, would interact with existing treatments. Our conclusions refer to 'cytochrome bc1 inhibitors' rather than to a specific compound in development.  
+
+<|ref|>sub_title<|/ref|><|det|>[[104, 417, 288, 433]]<|/det|>
+## Materials and Methods  
+
+<|ref|>text<|/ref|><|det|>[[102, 434, 835, 469]]<|/det|>
+The Materials & Methods section is well- detailed and allows for experimental reproducibility. However, there we have some comments:  
+
+<|ref|>text<|/ref|><|det|>[[102, 469, 890, 520]]<|/det|>
+- As in the Results section, avoid referring only to different experiments as Study A, B, etc. Instead, add descriptive headings such as "Intranasal Mice Challenge", "Intravenous Mice Challenge", or "High-Dose Aerosol Challenge."  
+
+<|ref|>text<|/ref|><|det|>[[102, 521, 880, 572]]<|/det|>
+- Each experimental model should consistently include the following information: Mouse strain, sex, and age; inoculum preparation, infection route, and dose; treatment duration and euthanasia time points.  
+
+<|ref|>text<|/ref|><|det|>[[102, 581, 895, 668]]<|/det|>
+We respectfully disagree with this suggestion. The experimental details for each study are already included in the Methods section, and adding them to the Results and figure legends would lead to unnecessary repetition. We have intentionally used this naming strategy to clearly distinguish studies from different institutions. Furthermore, many of the models have similar experimental parameters, making it difficult to differentiate them by name alone.  
+
+<|ref|>text<|/ref|><|det|>[[104, 678, 661, 695]]<|/det|>
+- Authors should provide in the ethical statement the end-point criteria.  
+
+<|ref|>text<|/ref|><|det|>[[104, 705, 714, 722]]<|/det|>
+We have now included the humane end points for the in vivo studies (pg 18).  
+
+<|ref|>text<|/ref|><|det|>[[102, 732, 870, 767]]<|/det|>
+- Supplementary tables and results should not be included in the Methods section but should be properly cited in the Results section.  
+
+<|ref|>text<|/ref|><|det|>[[102, 777, 892, 811]]<|/det|>
+Where possible, we have discussed supplementary tables and results in the Results section, however we disagree that these elements cannot be cited in the methods.  
+
+<|ref|>text<|/ref|><|det|>[[102, 821, 890, 891]]<|/det|>
+- There are several experimental methods referenced in the supplementary material but not appearing in the main text (e.g PK and tolerability in mouse, ADME assays, HepG2 cytotoxicity assay, mitochondrial toxicity assay Glu/Gal, Ames II Mutagenicity assay, mutant isolation and WGS). These should be briefly mentioned in the main text, such as: "Both JNJ-2901 and JNJ-4052
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[102, 80, 890, 115]]<|/det|>
+were evaluated for ADME and toxicity. All parameters supported further use of these compounds in subsequent experiments (Supplementary Table S2)".  
+
+<|ref|>text<|/ref|><|det|>[[104, 125, 661, 142]]<|/det|>
+We have now included this information in the Results section (pg 5- 6).  
+
+<|ref|>text<|/ref|><|det|>[[104, 153, 666, 187]]<|/det|>
+Additionally, we suggest to address the following minor comments: - L62: we would suggest classifying tuberculosis as a global pandemic.  
+
+<|ref|>text<|/ref|><|det|>[[105, 197, 395, 214]]<|/det|>
+We have updated this as suggested.  
+
+<|ref|>text<|/ref|><|det|>[[105, 225, 412, 241]]<|/det|>
+- L79: MoA abbreviation is not needed.  
+
+<|ref|>text<|/ref|><|det|>[[105, 252, 445, 268]]<|/det|>
+'MoA' is used elsewhere in the manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[104, 279, 881, 348]]<|/det|>
+- L87: the word "validated" to describe the inhibitor compound JNJ-2901 is confusing as the presented manuscript seems to be the validation of the compound as an effective treatment adjuvant. If this is not the case, please provide the reference where the compound is validated for its inhibition of the cytochrome bc1.  
+
+<|ref|>text<|/ref|><|det|>[[102, 358, 891, 392]]<|/det|>
+We have kept 'validated' as we can now present an additional manuscript (PMID:40191462), which outlines this validation.  
+
+<|ref|>text<|/ref|><|det|>[[102, 402, 888, 437]]<|/det|>
+- L91-96: we suggest to relocate this in the line 76, as the TB-PRACTECAL trial results are a good justification of linezolid-associated adverse effects when treating MDR-TB.  
+
+<|ref|>text<|/ref|><|det|>[[104, 448, 633, 464]]<|/det|>
+We have now moved this section to the Introduction as suggested.  
+
+<|ref|>text<|/ref|><|det|>[[104, 475, 567, 491]]<|/det|>
+- L116-122: please contextualize the results from this part.  
+
+<|ref|>text<|/ref|><|det|>[[104, 502, 461, 518]]<|/det|>
+We have expanded the results as requested.  
+
+<|ref|>text<|/ref|><|det|>[[104, 529, 692, 546]]<|/det|>
+- L127: move the figure citation to line 133. After the results are explained.  
+
+<|ref|>text<|/ref|><|det|>[[105, 557, 390, 572]]<|/det|>
+This has been moved as requested.  
+
+<|ref|>text<|/ref|><|det|>[[104, 583, 884, 617]]<|/det|>
+- L146: please provide the justification of why using Telacebec in this study and not JNJ-2901 and provide rationale about the drug regimen used for this experimental design.  
+
+<|ref|>text<|/ref|><|det|>[[104, 628, 894, 749]]<|/det|>
+This study was inspired by previous in vitro work demonstrating that the combination of bedaquiline, clofazimine, and telacebec resulted in rapid killing (PMID:27506290). We aimed to evaluate the in vivo efficacy of treatment regimens containing these compounds in combination with pyrazinamide, which has also been shown to enhance efficacy, potentially supporting treatment shortening (PMID:25622149). To allow comparison with previous work, we continued to use telacebec in this study rather than JNJ-2901. We have now clarified this justification in the results (pg 7) and discussion (pg 9).  
+
+<|ref|>text<|/ref|><|det|>[[104, 759, 879, 826]]<|/det|>
+- L164: as far as we understood, JNJ-2901 and Telacebec were also assessed for its bactericidal activity against clinical isolates in comparison to the H37Rv strain (Figure 2 B,C,D and E). Please provide a proper description of the compound used for this experiment and further discussion of this results in the main text.  
+
+<|ref|>text<|/ref|><|det|>[[104, 837, 894, 906]]<|/det|>
+At the time that this experiment was initiated, JNJ- 4052 was our lead molecule from our cytochrome bc1 drug discovery programme. This was the main reason it was chosen for the in vivo study outlined in Figure 3e. This compound provides a proof-of-principle for inhibition of cytochrome bc1, with other compounds with this MoA expected to have the same response as suggested by Figure 3b-d.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[102, 80, 870, 115]]<|/det|>
+- Figure 1: If the mice strain and experimental design are the same, consider re-structure Figure 1B and 1C in a unique figure.  
+
+<|ref|>text<|/ref|><|det|>[[102, 124, 875, 160]]<|/det|>
+We would prefer to keep these studies separate, enabling the reader to discriminate between the individual experiments.  
+
+<|ref|>text<|/ref|><|det|>[[102, 169, 893, 239]]<|/det|>
+- Figure 2: Please, consider a figure rearrangement. We suggest dividing Figure 2 in two. Section A on one side (resulting in Figure 2) and sections B, C, D, E and F on the other (resulting in Figure 3 A, B, C, D and E). Since the results represented refer to different experimental questions and require a results section for each one.  
+
+<|ref|>text<|/ref|><|det|>[[104, 248, 696, 266]]<|/det|>
+We agree, and have now split Figure 2 into two new figures as suggested.  
+
+<|ref|>sub_title<|/ref|><|det|>[[104, 276, 292, 292]]<|/det|>
+## Final Recommendation  
+
+<|ref|>text<|/ref|><|det|>[[104, 293, 883, 380]]<|/det|>
+This study presents valuable preclinical data on cytochrome bc1 inhibitors in TB treatment. However, we have significant concerns regarding the manuscript as it currently stands. It lacks a proper structure, clarity, and experimental justification, and the discussion is entirely insufficient. The manuscript needs to be totally rewritten to address these issues, ensuring that it achieves the required readability, scientific rigor, and overall impact to be considered suitable for publication.
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[120, 85, 424, 100]]<|/det|>
+## Response to reviewer's comments  
+
+<|ref|>text<|/ref|><|det|>[[118, 112, 844, 146]]<|/det|>
+We have addressed the final reviewer's comment in red below. We once again thank the reviewer's for their valuable comments.  
+
+<|ref|>text<|/ref|><|det|>[[120, 159, 430, 175]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[118, 193, 870, 279]]<|/det|>
+The revised manuscript of Clara Aguilar Pérez and colleagues has addressed all of my previous concerns and I am happy to see some of the recommendation on the revised discussion section as well as extended new data. The study represents an important progress on inhibition of respiration machinery in Mtb and has potential for the development of future therapeutics.  
+
+<|ref|>text<|/ref|><|det|>[[119, 314, 430, 330]]<|/det|>
+Reviewer #3 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[118, 348, 866, 469]]<|/det|>
+We would like to thank the authors for thoughtfully addressing our major suggestions regarding the manuscript. Our feedback primarily focused on improving its structure, clarity, and the discussion surrounding the experimental procedures and results. In particular, the authors have now effectively justified the differences observed when treating laboratory strains with different cytochrome bc1 inhibitors versus clinical strains. With these revisions, the manuscript now clearly highlights the significance of this investigation in advancing our understanding of tuberculosis treatment strategies.  
+
+<|ref|>text<|/ref|><|det|>[[117, 487, 861, 625]]<|/det|>
+This work underscores the potential of bc1 inhibitors as viable alternatives to fluoroquinolones and linezolid, which are often associated with adverse effects in current regimens for drug- resistant tuberculosis. Moreover, it contributes to the possibility of treatment shortening for drug- sensitive tuberculosis, with added potential against clinical strains. The experiments were performed using different, well- justified preclinical mouse models, which strengthens the translational relevance of the findings and supports further investigation of this promising asset. Based on these improvements and the scientific merit of the study, we believe the manuscript is now better suited for publication.  
+
+<|ref|>text<|/ref|><|det|>[[118, 643, 749, 659]]<|/det|>
+However, we would like the authors to address the following remaining issue:  
+
+<|ref|>text<|/ref|><|det|>[[116, 677, 877, 901]]<|/det|>
+Please add a paragraph on limitations to the discussion section. For example, the sample size of mice used in the experiments could be acknowledged as a limitation (even though we recognize that this is already mentioned on page 7). Additionally, regarding the authors' response that "no differences between sexes have been reported in relation to drug efficacy in TB mouse models," we would like to stress that, even if ethical requirements do not mandate the inclusion of both biological sexes, doing so is strongly advisable. There is robust evidence from both clinical and preclinical studies that biological sex influences TB epidemiology and treatment outcomes(for example, see: Idris, R., Infection 2025, https://doi.org/10.1007/s15010-024-02424-5; Dutta NK, Front Immunol, 2020, doi: 10.3389/fimmu.2020.01465; Tannenbaum, C., Nature, 2019, https://doi.org/10.1038/s41586-019-1657-6). These differences are driven by a combination of hormonal, genetic, and immunological factors, as well as potential differences in treatment adherence and broader social determinants of health. While we acknowledge that using female mice can be more
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[118, 83, 878, 222]]<|/det|>
+practical in terms of logistics and animal management- since males often display more aggressive behavior that can complicate experiments- there is now ample evidence showing that men and women often have different disease trajectories (not only for TB!) and responses to treatment. One persistent issue in drug development is that preclinical and clinical studies frequently fail to reflect the biological diversity found in real- world populations. Including both sexes in preclinical experiments wherever feasible would help improve the generalizability of the findings and ensure that results more accurately inform treatment strategies for all patients.  
+
+<|ref|>text<|/ref|><|det|>[[118, 234, 880, 354]]<|/det|>
+We thank the reviewer for their helpful suggestion. We have now added a paragraph to the discussion (pg. 10) outlining the limitations of our study and lessons for future work. Specifically, we address the use of female mice, as suggested, and also note the following points for improvement of future studies: (i) increasing statistical power by using larger animal group sizes, (ii) incorporating monotherapy arms to enable regimen deconvolution, and (iii) including clinical isolates to better reflect the diversity and relevance of circulating M. tuberculosis strains.  
+
+<|ref|>text<|/ref|><|det|>[[119, 383, 880, 435]]<|/det|>
+In summary, we encourage the authors to carefully consider any additional limitations of their study and clearly state them, as this will add valuable context and transparency. Once this is done, the manuscript can be endorsed for publication.  
+
+<|ref|>text<|/ref|><|det|>[[120, 465, 430, 481]]<|/det|>
+Reviewer #4 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[119, 494, 880, 545]]<|/det|>
+I co- reviewed this manuscript with one of the reviewers who provided the listed reports. This is part of the Nature Communications initiative to facilitate training in peer review and to provide appropriate recognition for Early Career Researchers who co- review manuscripts.
+
+<--- Page Split --->

peer_reviews/12bee38be00ca6148fc0c8a69c26bf2450d3af9d3ae7dfb34f177755640c2e32/supplementary_0_Peer Review File/images_list.json

@@ -0,0 +1,40 @@
+[
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_0.jpg",
+    "caption": "Figure L1: Comparison of biological triplicates for the 43 validation variants grown in SD-\\(Ura + 2\\%\\) glucose. Experiments show a strong a correlation between all 3 repeats.",
+    "footnote": [],
+    "bbox": [],
+    "page_idx": 0
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_1.jpg",
+    "caption": "Figure L2: Fitness analysis for yeast cells expressing the 43-variant validation set grown in YPD. (a) The 43 validation variants were grown in YPD in duplicates and tracked for OD as a function of time (circles). For each variant the growth data are fitted (blue lines) by a classic growth curve (see Supplementary Information). (b) The rates of growth for each variant are plotted as a scatter plot pair for both repeats.",
+    "footnote": [],
+    "bbox": [
+      [
+        123,
+        298,
+        861,
+        581
+      ]
+    ],
+    "page_idx": 5
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_2.jpg",
+    "caption": "Figure L3:Cross correlation data for 43-variant validation experiments plotted with error-bars. (a) yeast cross-correlation data. (b-c) The \\(2\\%\\) glucose yeast expression data plotted as a function of CHO (b) and HeLa (c) expression data. Error-bars were computed using standard-error analysis carried out on mean flow cytometry fluorescence measurements obtained from three or four biological repeats – depending on data set.",
+    "footnote": [],
+    "bbox": [
+      [
+        123,
+        383,
+        863,
+        607
+      ]
+    ],
+    "page_idx": 7
+  }
+]

peer_reviews/12bee38be00ca6148fc0c8a69c26bf2450d3af9d3ae7dfb34f177755640c2e32/supplementary_0_Peer Review File/supplementary_0_Peer Review File.mmd

@@ -0,0 +1,460 @@
+
+# nature portfolio  
+
+Peer Review File  
+
+A universal system for boosting gene expression in Eukaryotic cell- lines  
+
+![](images/Figure_unknown_0.jpg)
+  
+
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+
+Reviewers' Comments:  
+
+Reviewer #1:  
+
+Remarks to the Author:  
+
+In this study, the authors introduce an algorithm designed to predict regulatory sequences that modulate gene expression in eukaryotic cells. To achieve this, they identified 41 Transcription Factor Binding Sites (TFBS) from various organisms, including S. cerevisiae, S. pombe, D. melanogaster S2 cells, and murine ES cells. These TFBS (or motifs) were used to construct a library of nearly 190,000 variants, each containing up to three different motifs and their variants. Using a combination of machine learning and oligo library analysis in yeast, the authors aimed to identify specific motifs capable of enhancing or attenuating gene expression. These motifs were then assessed in two mammalian cell lines.  
+
+Overall, the study addresses a topic of general interest: fine tunability of gene expression for optimal output production. However, the manuscript suffers from issues related to clarity, methodology transparency, and validation, ultimately raising concerns about the rigour and robustness of the study. Detailed comments and suggestions are provided below.  
+
+1. The TFBS used in this study are for endogenous TFs, and some correlate with both activation and repression activity. This may depend on different cell types, cell states and growth conditions, potentially generating high variability in the promoter's behavior. The authors should validate the identified sequences with a significant number of biological replicates (at least 3), at different cell passages and possibly different culturing conditions (e.g. different cell densities, different media composition, etc).  
+
+2. The authors should ensure that the cells carrying a specific barcode indeed harbour the intended promoter variant, especially considering the complexity of the designed library. Thus, one significant gap in the methodology is the absence of verification steps after sorting, ideally conducted after the final PCR step just before Next-Generation Sequencing (NGS). This validation step could take various forms, even only Sanger sequencing of a select subset of variants (not only 1!). This additional verification would provide confidence in the accuracy of the cell content and the presence of the intended promoter variants.  
+
+3. Figure 2c should ideally show a relatively tight cluster of data points resembling a straight line with some expected noise and a few outliers, reflecting the inherent variability in biological systems and NGS data. However, the observed data points in this plot appear scattered, indicating a significant divergence from expected behavior. To mitigate this concern and validate the Sort-Seq values, it is advisable to conduct an extensive validation experiment. This experiment could involve the measurement of fluorescence or RNA-seq for individual variants, followed by a correlation analysis with the NGS-based fluorescence (FL mean) values. If this validation does not yield a strong correlation, it could raise doubts about the reliability of the entire experimental approach and data interpretation. Therefore, addressing these issues through thorough verification and validation steps is crucial to enhance the study's credibility and confidence in the results.  
+
+4. Increased gene expression, can correlate with reduced host organism fitness (DOI: https://doi.org/10.1038/nmeth.3339) and unwanted competition for intracellular resources (DOI: https://doi.org/10.1038/s41467-020-18392-x). If the final goal is to characterize new promoters for bioproduction, the authors should show that they do not impact the fitness of the host cells.  
+
+5. Both introduction and discussion need more supporting citations. Please add evidences to support all claims. Additionally, some claims seem only partially true or imprecise:  
+
+a. Line 47: there are studies that addressed the same problem (e.g. DOI: https://doi.org/10.1128/aem.00939-22). These should be cited and discussed. 
+b. Lines 81-85: these claims should be either supported by citations or toned down. There are many well characterized constitutive promoters active across eukaryotic species (CMV, EF1a, PGK, etc). The authors should motivate why their design is superior to existing solutions.  
+
+6. Some experimental choices and pipeline steps are unclear:  
+
+a. How were the 41 TFBS selected, especially since many are not common across the 4 eukaryotic cell types evaluated?  
+
+b. Line 128: how was the threshold of \(>73\%\) determined and why?  
+
+c. In the method section, describe how the total of 1600 million NGS reads mentioned is processed and the steps involved. Why only \(25\%\) of NGS reads were selected? What is the potential impact on the results?  
+
+7. Some of the conclusions drawn by the authors in the results section are questionable. Can the
+
+<--- Page Split --->
+
+authors comment on the following points?  
+
+a. In Fig. 2a the authors claim that there is a clear correlation between the median of the mean fluorescence and the motif: this is not clear from this graphic and it could be pure randomness.  
+b. Line 236: the observed correlation indicates that the distribution is mainly not described by the model.  
+c. Line 269-271 and Fig. 3d: there is no clear correlation between an increase in mixed letters and increase in model performance. The values of 10, 26, 36, 16, 23, 18% of correlation could be more or less random numbers.  
+8. The manuscript presentation is unclear in some parts, especially in the results and discussion sections preventing full understanding of the study:  
+a. The term "mean FL" is a function of the NGS reads. This is confusing as "mean FL" should intuitively be the mean of the measured fluorescence only.  
+b. Is "mean fluorescence" (Fig. 2a) same as "mean FL" (Fig. 2c and others)?  
+c. Large parts of the results section suffer from a lack of clarity, for example lines 168-190. Please work on improving the clarity of the entire section.  
+d. The structure of the discussion is confusing: start with a brief summary of the work, then draw a conclusion and hypothesize the next steps, only to return to the summary in the middle of the second paragraph. Consider restructuring.  
+9. There are several overstatements in the manuscript that should be toned down. E.g.:  
+a. Line 106 "pan-organism": the study focuses on one yeast strain and two mammalian cell lines only.  
+b. Line 112 "different murine tissues": this is inconsistent with the caption of Supplementary Figure 1 where only mouse embryonic stem cells are mentioned.  
+10. Some scientific basics are lacking:  
+a. space between number and units  
+b. organism in italic  
+c. missing descriptions in the methods (see comment 6c)  
+d. missing code and data online: will these be provided after publication?  
+e. the unit of fluorescence is never stated, is it arbitrary units (au)?  
+f. number of biological replicates should be specified. Especially Figure 6 does not have error bars: was validation run on one biological replicate only? If this is the case, it is scientifically inaccurate to draw any conclusion.  
+
+Minor comments:  
+
+1. Line 49: "whose promoter is capable of transcribing" is wrong. A promoter does not transcribe, the polymerase does.  
+2. Line 102: "reliable prediction"  
+3. Line 106: "constitutive" instead of non-inducible  
+4. Line 156: "Fig. 1f"  
+5. Line 802: what is a PPM?  
+6. Line 140: Fig. 1c?  
+7. Line 185: which type of correlation is it?  
+8. Line 422: "for a weak and a strong promoter", only one variant tested for each  
+9. Line 430: "circuits" (Remove "bio-")  
+10. Line 869: "of" "s"?  
+11. Lines 780-781: lasers wavelength?  
+12. Discussion: The name UNILIB is introduced only at this point, why?  
+13. Figures:  
+a. Increase the font size in all figures, as they appear too small for easy readability  
+b. 1f: the vertical line should be at 40, it is at 30 though  
+c. 2d: no error bars  
+d. 3e: labels can be put under x axis instead of colors  
+e. 3d: if a linear correlation (Pearson) is assumed, add the line to the plot  
+f. 4a: error bars are at null level?  
+g. 4b (and following): no tick marks in the right plots  
+h. 5f: error bars  
+i. 6: add axis label, what type of data is shown?
+
+<--- Page Split --->
+
+Reviewer #2:  
+
+Remarks to the Author:  
+
+In this manuscript from Vaknin et al, the authors develop a library of synthetic promoters by merging a target core promoter and upstream TF binding sites, to enhance protein expression in eukaryotic cells. This is not the first time synthetic promoters are designed by this approach and previous work has shown that is possible to engineer mammalian cells based on mining of TFBS and core promoter engineering (e.g. Johari et al. 2019 and others).  
+
+The manuscript tries to go beyond the state of art by building a system functional in more than one chassis.  
+
+I have a few questions, mainly on the experimental side, as I am not a computational person. It is not fully clear to me what the authors mean with the term "motif" and it could be useful to define this.  
+
+It is not fully clear also how the 41 motifs were identified and if this was validated by using more than one TF data base.  
+
+In figure 1e it is not fully clear why authors have not taken population 2 and re- bin it in order to capture more variants and potentially more interesting candidates with higher expression. It could have been beneficial to do so here.  
+
+In figure 6, for CHO cell experiments, the authors mentioned they used a BFP for fluorescence normalisation. However, it is known from literature that resource competition can impact normalisation (see Frei et al, 2020. Jones et al, 2020).  
+
+Can the authors shown the expression levels of the CHO library but with no normalisation and check if that improves the results? Caption of figure 6 should also more clearly describe what the figures shows.  
+
+Minor comments pertain to several typos present in the manuscript like Eukaryotic that should be lower case;  
+
+Figures are called at time with capital letter and at times with lower case.  
+
+In conclusion I suggest the work to be published once these questions have been addressed.
+
+<--- Page Split --->
+
+Reviewer #1 (Remarks to the Author):  
+
+In this study, the authors introduce an algorithm designed to predict regulatory sequences that modulate gene expression in eukaryotic cells. To achieve this, they identified 41 Transcription Factor Binding Sites (TFBS) from various organisms, including S. cerevisiae, S. pombe, D. melanogaster S2 cells, and murine ES cells. These TFBS (or motifs) were used to construct a library of nearly 190,000 variants, each containing up to three different motifs and their variants. Using a combination of machine learning and oligo library analysis in yeast, the authors aimed to identify specific motifs capable of enhancing or attenuating gene expression. These motifs were then assessed in two mammalian cell lines. Overall, the study addresses a topic of general interest: fine tunability of gene expression for optimal output production. However, the manuscript suffers from issues related to clarity, methodology transparency, and validation, ultimately raising concerns about the rigour and robustness of the study. Detailed comments and suggestions are provided below.  
+
+1. The TFBS used in this study are for endogenous TFs, and some correlate with both activation and repression activity. This may depend on different cell types, cell states and growth conditions, potentially generating high variability in the promoter's behavior. The authors should validate the identified sequences with a significant number of biological replicates (at least 3), at different cell passages and possibly different culturing conditions (e.g. different cell densities, different media composition, etc).  
+
+We thank the reviewer for this comment, and the opportunity to improve our work. First, we would like to note that only 27 of the 41 motifs were validated transcription-factor binding sites. The rest of the motifs were unknown, and not attributed to a particular regulatory process. The motifs were found to be conserved across various Eukaryotic lineages in a HT-SELEX assay carried out by Jussi Taipale's group (published in DOI: 10.1038/nbt.4138). As a result of the UNILIB experiment, we were able to characterize 5 of the unknown motifs as either up- or down-regulating. Some of these motifs are weak and require multiple sites or combinations of sites to show a measurable regulatory effect.  
+
+Second, with regards to the function of the motifs in different cell- growth conditions or cell types, there indeed could be a variation. Consequently, we created 43 unseen synthetic URS variants. We used 32 variants, which sampled 23 of the motifs for which a significant up or down- regulatory behavior was identified, for the results depicted in Fig. 5, and an additional 11 variants were used for the validation of the machine- learning model in Fig. 3. Altogether the 43 validation variants sample 31 of the 42 motifs including all the motifs that were found to be significant.  
+
+However, given the reviewer's critical comment, we decided to use these variants to expand our validation set to further strengthen our conclusions. The revised manuscript now includes two new figures: Fig. 6 and Fig. 7, which replace the original Fig. 6. New Fig. 6 details the results of 7 separate experimental measurements carried out in biological triplicates as follows:  
+
+1. Yeast: SD-Ura+2% glucose, \(30^{\circ}\mathrm{C}\) , weak core promoter (mCore), no additives, 43 variants. 
+2. Yeast: SD-Ura+2% glycerol, \(30^{\circ}\mathrm{C}\) , weak core promoter (mCore), no additives, 43 variants. 
+3. Yeast: SD-Ura+2% glucose, \(39^{\circ}\mathrm{C}\) , weak core promoter (mCore), no additives, 43 variants. 
+4. Yeast: SD-Ura+2% glucose, \(30^{\circ}\mathrm{C}\) , weak core promoter (mCore), 1M NaCl, 43 variants. 
+5. Yeast: SD-Ura+2% glucose, \(30^{\circ}\mathrm{C}\) , strong core promoter (FEC-mCore), no additives, 20 variants. 
+6. CHO/K1 cells grown in F12 medium, 32 variants. 
+7. HeLa cells grown in DMEM high glucose medium, 32 variants.  
+
+The results for all the yeast experiments, whether as biological replicates or in different culturing conditions, correlate strongly with one another in a statistically significant fashion (see below Figs. L1 in addition to Fig. 6 and Supplemental Fig. 6 in the revised manuscript). These results support a conserved function in yeast independent of growth conditions. This is consistent with the previous findings of (Keren et al., Mol.
+
+<--- Page Split --->
+
+Sys. Biol., 9:701 (2013)) who tested the activities of \(\sim 900\) S. cerevisiae promoters in 10 different conditions and found no significant difference in activities across the panel of growth conditions used. In addition, the experimental data also shows a statistically significant correlation between the mammalian cell measurements and the yeast measurements, which we showed also in the original manuscript. Finally, we also improved our modelling scheme by extending the machine- learning model developed in Fig. 3 using the additive model described in Fig. 5 to create a hybrid model, which improved performance on all datasets. The new hybrid model now predicts HeLa and CHO cells regulatory activity in a statistically significant fashion (see new Fig. 7).  
+
+The revised validation segment of our manuscript now allows us to make an improved argument that the synthetic URS sequences that we created can generate a regulatory function that is, for the most- part, independent of cell- type and growth conditions, making these sequences a potentially widely applicable tool.  
+
+![](images/Figure_unknown_1.jpg)
+
+<center>Figure L1: Comparison of biological triplicates for the 43 validation variants grown in SD-\(Ura + 2\%\) glucose. Experiments show a strong a correlation between all 3 repeats. </center>  
+
+2. The authors should ensure that the cells carrying a specific barcode indeed harbour the intended promoter variant, especially considering the complexity of the designed library. Thus, one significant gap in the methodology is the absence of verification steps after sorting, ideally conducted after the final PCR step just before Next-Generation Sequencing (NGS). This validation step could take various forms, even only Sanger sequencing of a select subset of variants (not only 1!). This additional verification would provide confidence in the accuracy of the cell content and the presence of the intended promoter variants.  
+
+In brief, due to their short length (bc+sURS+core- promoter) our variants were fully sequenced in the NGS. While we apologize for not including this data in the manuscript, as part of our validation, we screened for all full- length reads in the library (i.e. reads that yielded a full length of over 200 bp that include the barcodes, sURS, and primer regions). We found approx. 400M such reads, and they all perfectly matched the barcodes, motifs, and primer regions. We then quantified the correctness of the downstream sequence to the barcode with the intended design and found that per position \(>99\%\) match the design. In addition, we designed the barcodes, such that the Hamming distance between each two is at least 3, which maximized our ability to identify variants. We thank the reviewer for this important comment. In response to the comment, we have added a comment under the Methods sub- section "NGS data processing and read normalization" to clarify this issue.  
+
+3. Figure 2c should ideally show a relatively tight cluster of data points resembling a straight line with some expected noise and a few outliers, reflecting the inherent variability in biological systems and NGS data.
+
+<--- Page Split --->
+
+However, the observed data points in this plot appear scattered, indicating a significant divergence from expected behavior. To mitigate this concern and validate the Sort- Seq values, it is advisable to conduct an extensive validation experiment. This experiment could involve the measurement of fluorescence or RNA- seq for individual variants, followed by a correlation analysis with the NGS- based fluorescence (FL mean) values. If this validation does not yield a strong correlation, it could raise doubts about the reliability of the entire experimental approach and data interpretation. Therefore, addressing these issues through thorough verification and validation steps is crucial to enhance the study's credibility and confidence in the results.  
+
+We are aware of the lack of correlation observed in our OL measurements, which we view as a feature of the methodology rather than an artefact. The data used to compose panel 2c consisted of OL variants that were encoded with two barcodes. The lack of correlation observed for most variant pairs was due to three main reasons:  
+
+1. In SORT-seq experiments, it is impossible to tell how many reads came from a single sorted cell. For example, one cell can yield several hundred reads or a single read in the final count. This can lead to a lack of correlation if the number of cells sorted per a particular variant is small as in this case. We remind the reviewer that our library was unprecedented in size ( \(\sim 200\mathrm{K}\) mixed-base variants), and thus the number of cells collected for each variant was small. 
+2. In SORT-seq experiments, it is impossible to tell whether cells die after sorting or divide. In such a case, a single cell may also yield more read events in comparison to another cell, though the original sort had them at an equal weight. 
+3. Variants which do not encode a particular regulatory function are more susceptible to transcriptional variation. The strongly up- and down-regulating variants (top and bottom corners of the plot respectively) are strongly correlated as expected from a fully functional regulatory element.  
+
+The analysis shown in Figure 2 and referred to by the reviewer was, therefore, used as impetus for the quantitative analysis that was done on the 2,435- variant subset that were encoded with 22 barcodes. We reasoned that the sources for the lack of correlation discussed in the points above will be minimized by analyzing multiple barcodes on the same variants. For example, identifying all 22 barcodes immediately implies that at the very least we screened 22 separate fluorescent yeast cells. Consequently, we reasoned that such a subset will lead to a more robust statistical analysis, even though a given pair of barcodes may not be as correlated as we would like.  
+
+The analysis carried out in Figure 3 shows that the choice of unifying the 22 barcodes is statistically robust as evident by the agreement with the de Boer machine- learning model that was trained using independent experimental data. In addition, the validation set of Figures 5- 7 and the agreement with the models that were derived from the experimental data obtained from the 2,435- variant subset essentially shows that both the SORT- seq experiment and the analysis were robust. Consequently, we believe that we have provided sufficient proof in the manuscript both computationally and experimentally that our analysis and experimental approach were robust, and no additional validation is necessary.  
+
+4. Increased gene expression, can correlate with reduced host organism fitness (DOI: https://doi.org/10.1038/nmeth.3339) and unwanted competition for intracellular resources (DOI: https://doi.org/10.1038/s41467-020-18392-x). If the final goal is to characterize new promoters for bioproduction, the authors should show that they do not impact the fitness of the host cells.  
+
+We thank the reviewer for this comment. In response to the comment, we carried out fitness analysis on the 32 validation variants. The results are shown in Figure L2 and have been added to the Supplementary Information as Supplementary Fig. 6. In our experiments, we did not detect any effect on fitness as a result of our synthetic upstream regulatory sequences and mentioned that in the "Boosted expression level translates from yeast and mammalian cells" Results subsection of the revised manuscript. Specifically, we grew the 43 validation variants in YPD and tracked their growth via OD measurement as function of time
+
+<--- Page Split --->
+
+(Fig. L2a - circles). For each strain we then fitted the OD measurements (Fig. L2a - blue line) with the following model for exponential growth:  
+
+\[OD_{600}(t) = C + \frac{L}{(1 + e^{-k(t - t_0)})}\]  
+
+Where \(C\) is background OD levels, \(L\) is the max OD, \(k\) is the growth rate, and \(t_0\) corresponds to the lag time (i.e. time at which the culture reaches OD of L/2). Using this model, we extracted the growth rate for each strain and for both repeats. We plotted in Fig. L2b the different fitted growth rates \((k)\) for both repeats that were measured for each variant. The results show that the growth rates for all experiments was found to be within a narrow range of \(\sim 0.4 - 1.3\) (1/hr) without any significant correlation between duplicates. The lack of correlation between duplicates and narrow range of growth rates provides strong evidence that our variants do not affect the fitness of the yeast cells.  
+
+![](images/Figure_unknown_2.jpg)
+
+<center>Figure L2: Fitness analysis for yeast cells expressing the 43-variant validation set grown in YPD. (a) The 43 validation variants were grown in YPD in duplicates and tracked for OD as a function of time (circles). For each variant the growth data are fitted (blue lines) by a classic growth curve (see Supplementary Information). (b) The rates of growth for each variant are plotted as a scatter plot pair for both repeats. </center>  
+
+5. Both introduction and discussion need more supporting citations. Please add evidences to support all claims. Additionally, some claims seem only partially true or imprecise: a. Line 47: there are studies that addressed the same problem (e.g. DOI: https://doi.org/10.1128/aem.00939-22). These should be cited and discussed. b. Lines 81-85: these claims should be either supported by citations or toned down. There are many well characterized constitutive promoters active across eukaryotic species (CMV, EF1a, PGK, etc). The authors should motivate why their design is superior to existing solutions.  
+
+We thank the reviewer for providing the additional reference, and for pointing out misstatements in the Introduction. In response to the comment, we added this and other references and rephrased the text in the Introduction and Discussion sections accordingly. Please see the marked up version of the revised manuscript for the details.  
+
+6. Some experimental choices and pipeline steps are unclear:
+
+<--- Page Split --->
+
+a. How were the 41 TFBS selected, especially since many are not common across the 4 eukaryotic cell types evaluated?  
+
+We apologize that these choices were unclear. In brief, the motifs were chosen as part of a past collaboration with Jussi Taipale's group. At the time, the Taipale group had a study (later published in DOI: 10.1038/nbt.4138) that was focused on developing a protein activity assay specifically for DNA- binding TFs in cell and tissue extracts. His team was able to identify strong and enriched TFBS motifs using this assay, including the 41 TFBSs that we selected for our research. Those and other motifs were discovered in numerous organisms, including two types of yeast, different tissues from mice, bacteria, and Drosophila S2 cells. Even though some of the motifs were not found to be common/conserved across these organisms, we nevertheless chose to characterize them due to similarities to well- characterized protein families in higher eukaryotes (e.g. the bHLH motif). For a detailed explanation as to how and why we chose these 41 motifs, please see the new Methods subsection titled "Motif selection and encoding into OL. Based on the study's data, we selected the 41 enriched motifs according to the following criteria:  
+
+1. Different motif types: 8 organism-shared motifs, 5 mice tissue-shared motifs, 14 unshared motifs (unique to an organism), and 14 unknown motifs (with unknown regulatory function). 
+2. Known/unknown regulatory function: 27 of the selected motifs have known regulatory function, and those were anticipated to be the control motifs. Characterizing the regulatory function of the 14 unknown motifs in yeast was one of the stated goals of the study.  
+
+b. Line 128: how was the threshold of \(>73\%\) determined and why?  
+
+We thank the reviewer for this question and apologize for not including the rationale for our choice in the original manuscript. In response to the comment, we have added a new Methods subsection entitled "Motif selection and encoding into OL", where the rationale is now described. In brief, K and M substitutions were based on the percentage calculation of the respective G/T and A/C occurrences, in each position of the motif according to the PFM data given to us by the Taipale group (later published in DOI: 10.1038/nbt.4138). Positions within the motif, with dominant percentages, were replaced by either K or M in the final design. \(70\%\) threshold was set to determine the K/M substitution, but the actual calculated threshold was higher at \(73\%\) , as specified in the Results section.  
+
+c. In the method section, describe how the total of 1600 million NGS reads mentioned is processed and the steps involved. Why only \(25\%\) of NGS reads were selected? What is the potential impact on the results?  
+
+We thank the reviewer for this question. In response to the comment, we have expanded the Methods subsection titled "NGS data processing and read normalization", specifying the processing steps accordingly. Given the size of the UNILIB library, we opted to use Illumina's NovaSeq for its ability to generate as much as 1B reads per run. We used Illumina's S1 kit, and despite Illumina's reliability claims, indeed only \(25\%\) of the reads were found to be correct in both the NovaSeq runs that were made. Neither we nor the Genomics Center at the Weizmann Institute are certain about the underlying cause for this low fidelity, which may have been due to the choice of primers. Irrespective of that, the center offered to re- sequence the library, ultimately allowing us to extract 400M correct reads. This number of reads proved to be sufficient for the analysis in our study, which was based on the machine- learning modelling.  
+
+As discussed above, our analysis showed that the use of two barcodes per variant was insufficient for proper quantitative analysis, as such a low number introduced a large statistical uncertainty regarding the actual number of sorted cells. On the flip side, the same analysis showed that 22 barcodes per variant significantly reduced this uncertainty. In the former case, increasing the number of reads would not have affected the cell number uncertainty, while in the latter the 22 barcodes ensured that we have a sufficient amount of reads for every one of the variants of the 2,435- variant subset even with only 400M correct reads extracted in total. Consequently, the \(25\%\) extraction rate did not impact our conclusions or results of the downstream analysis.
+
+<--- Page Split --->
+
+7. Some of the conclusions drawn by the authors in the results section are questionable. Can the authors comment on the following points?  
+
+a. In Fig. 2a the authors claim that there is a clear correlation between the median of the mean fluorescence and the motif: this is not clear from this graphic and it could be pure randomness.  
+
+We thank the reviewer for this comment and apologize for not assessing the correlation in the original manuscript. In response to the comment, to assess the correlation between the median of mean fluorescent expression levels and motifs, we conducted a Wilcoxon rank-sum test by comparing the mean fluorescent expression levels of the group of variants containing each motif to the group of variants containing motifs ranked at least 5 motifs away. To correct for multiple tests, we applied the Benjamini-Hochberg procedure with an FDR threshold of 0.1. This new analysis revealed significant differences for 36 out of 42 motifs (p- value<0.05), where the 6 non- significant motifs are in motifs ordered 15- 20 in decreasing order, which is expected since their values are distributed around the center of the mean- fluorescence distribution. We added this new statistical significance analysis to the Results section.  
+
+b. Line 236: the observed correlation indicates that the distribution is mainly not described by the model.  
+
+We thank the reviewer for this comment. In various biological problems, such as inference of protein- DNA binding preferences based on protein- binding- microarray data, correlation values are within 0.4- 0.6 (10.1109/TCBB.2019.2947461). Moreover, the correlation of 0.45 was achieved by the all- data model (ADM), which we showed is inferior to the AMM and MBO models. In addition, this value is calculated on a test set of variants, where most have only 2 barcodes, which implies the lower quality of their mean- fluorescence measurements compared to variants with 22 barcodes. One of the key insights from our study was that variants supported by 22 barcodes, as opposed to 2 barcodes, yielded more accurate mean- fluorescence measurements (due to the inclusion of more cells carrying the variant). When the MBO model was exclusively trained and tested on datasets of variants with 22 barcodes, the obtained Pearson correlation was 0.61. This suggests that using more accurate training and test sets results in a higher observed correlation, which better explains the variability in the data.  
+
+c. Line 269-271 and Fig. 3d: there is no clear correlation between an increase in mixed letters and increase in model performance. The values of 10, 26, 36, 16, 23, 18% of correlation could be more or less random numbers.  
+
+We thank the reviewer for pointing this unclear part of our manuscript. We acknowledge that the reported correlation values may not achieve statistical significance. In response to this comment, we removed Fig. 3d and the corresponding text describing the analyses and results from the Results section.  
+
+8. The manuscript presentation is unclear in some parts, especially in the results and discussion sections preventing full understanding of the study:  
+
+a. The term "mean FL" is a function of the NGS reads. This is confusing as "mean FL" should intuitively be the mean of the measured fluorescence only.  
+
+We thank the reviewer for this comment and have changed all "mean FL" labels for the NGS data to the one used by (Sharon, E. et al. Nat Biotechnol 30, 521-530 (2012)) in a similar oligo library experiment: "Expression (A.U.)". All relevant labels in the figures and text have been changed to Expression (A.U.) in the revised manuscript.  
+
+b. Is "mean fluorescence" (Fig. 2a) same as "mean FL" (Fig. 2c and others)? We thank the reviewer for finding this minor error, and changed to Expression (A.U.).
+
+<--- Page Split --->
+
+c. Large parts of the results section suffer from a lack of clarity, for example lines 168-190. Please work on improving the clarity of the entire section.  
+
+We thank the reviewer for this comment and have clarified the Results, including the subsection titled "Variants manifest a broad range of regulatory behavior". Please see the marked-up version of the revised manuscript for the changes.  
+
+d. The structure of the discussion is confusing: start with a brief summary of the work, then draw a conclusion and hypothesize the next steps, only to return to the summary in the middle of the second paragraph. Consider restructuring.  
+
+We thank the reviewer for this comment. We decided to leave the structure of the Discussion as is with minor modifications. We opted for this option so we can highlight at the beginning of the section the design algorithm that was developed, which we believe is the most important achievement of this work.  
+
+9. There are several overstatements in the manuscript that should be toned down. E.g.: a. Line 106 "pan-organism": the study focuses on one yeast strain and two mammalian cell lines only.  
+
+In response to the comment, we deleted the term pan-organism from the Introduction.  
+
+b. Line 112 "different murine tissues": this is inconsistent with the caption of Supplementary Figure 1 where only mouse embryonic stem cells are mentioned.  
+
+In response to the comment, we corrected the error and replaced the term "different murine tissues" by mouse (ES cells and different tissues).  
+
+10. Some scientific basics are lacking:  
+
+a. space between number and units  
+
+Errors were corrected.  
+
+b. organism in italic  
+
+Errors were corrected.  
+
+c. missing descriptions in the methods (see comment 6c)  
+
+All missing methods were added. Please see responses above.  
+
+d. missing code and data online: will these be provided after publication?  
+
+All data and code were uploaded. Please see data and code availability statements in the revised manuscript.  
+
+e. the unit of fluorescence is never stated, is it arbitrary units (au)?  
+
+The mean FL was replaced by the consensus Expression (A.U.) for all mean level expression level measurements.  
+
+f. number of biological replicates should be specified. Especially Figure 6 does not have error bars: was validation run on one biological replicate only? If this is the case, it is scientifically inaccurate to draw any conclusion.  
+
+We thank the reviewer for this comment and would like to clarify our choice of plotting. All validation experiments were conducted in (at least) biological triplicates as discussed above (see Fig. L1 for example). In response to Comment 1, we have substantially increased the volume of the validation experiments, and consequently split the original Fig. 6 in the original manuscript to a new Fig. 6 depicting only experimental data, and a new Figure 7 depicting the modelling analysis. Given the quantity of data, we opted to omit the error-bars from some of the plots for aesthetic reasons (see for example Fig. L3 which depicts Fig. 6a-c
+
+<--- Page Split --->
+
+with error bars). Since we recognize the need to visually and numerically report the variance in the validation measurements, we have done the following:  
+
+- The variation for the various validation datasets across the repeats can be assessed using new Supplementary Figure 6, and via the correlation heatmap plots of new Fig. 6 (panels d and e). Note, that the autocorrelation between the various repeats of the yeast experiments does not yield a higher Pearson correlation coefficient, as compared with the cross-correlation of the mean fluorescent expression levels of the various conditions. This result provides statistically significant experimental evidence that the different yeast growth conditions do not alter the expression of our variants, and in our opinion is depicted more convincingly via the mode of plotting that we chose to use in the revised manuscript.- We plotted Fig.6a with error bars to convey directly the variance in the cross-correlation analysis. This panel is identical to the one shown in Fig. L3a.- Fig. 6b-c are depicted without error-bars. Here, the purpose is to highlight the fact that out of the dataset 24 variants behaved similarly to yeast (blue), while 8 were uncorrelated in CHO (red). We have added the panel b and c to supplementary fig.6 as panel g and h.- We plotted the error-bars for the both the yeast-glucose \(2\%\) and HeLa-blue variants in panel b and c of Fig. 7.  
+
+![PLACEHOLDER_11_0]
+
+<center>Figure L3:Cross correlation data for 43-variant validation experiments plotted with error-bars. (a) yeast cross-correlation data. (b-c) The \(2\%\) glucose yeast expression data plotted as a function of CHO (b) and HeLa (c) expression data. Error-bars were computed using standard-error analysis carried out on mean flow cytometry fluorescence measurements obtained from three or four biological repeats – depending on data set. </center>  
+
+Minor comments:  
+
+1. Line 49: "whose promoter is capable of transcribing" is wrong. A promoter does not transcribe, the polymerase does.  
+
+We changed the statement to "whose promoter is capable of initiating transcription"  
+
+2. Line 102: "reliable prediction"  
+
+We removed the word prediction.  
+
+3. Line 106: "constitutive" instead of non-inducible  
+
+We retrained the word non-inducible.  
+
+4. Line 156: "Fig. 1f"
+
+<--- Page Split --->
+
+We corrected the typo.  
+
+5. Line 802: what is a PPM?  
+
+We corrected PPM to PWM - position weighted matrix.  
+
+6. Line 140: Fig. 1c?  
+
+We corrected the typo.  
+
+7. Line 185: which type of correlation is it?  
+
+We added the word "Pearson".  
+
+8. Line 422: "for a weak and a strong promoter", only one variant tested for each  
+
+We changed the phrase in accordance with the reviewer's suggestion.  
+
+9. Line 430: "circuits" (Remove "bio- ")  
+
+We removed the word bio.  
+
+10. Line 869: "of""s"?!  
+
+We corrected the typo.  
+
+11. Lines 780-781: lasers wavelength?  
+
+We added MacsQuant laser wavelength.  
+
+12. Discussion: The name UNILIB is introduced only at this point, why?  
+
+We chose to name the completed design algorithm UNILIB, and name it only after we demonstrated that it worked via the validation experiments.  
+
+13. Figures:  
+
+a. Increase the font size in all figures, as they appear too small for easy readability  
+
+We corrected where possible.  
+
+b. 1f: the vertical line should be at 40, it is at 30 though  
+
+We corrected this error.  
+
+c. 2d: no error bars  
+
+There are no error bars for the medians of the mean fluorescent expression level distributions.  
+
+d. 3e: labels can be put under x axis instead of colors  
+
+We improved the figure accordingly.  
+
+e. 3d: if a linear correlation (Pearson) is assumed, add the line to the plot  
+
+We removed old Fig. 3d.  
+
+f. 4a: error bars are at null level?  
+
+There are no error bars for the large colored bars as they represent the p- value estimate from comparing two distributions as shown in the inset.  
+
+g. 4b (and following): no tick marks in the right plots  
+
+We corrected this error.  
+
+h. 5f: error bars  
+
+We opted to present the actual measurements as circles instead of error bars to convey the distribution of the measured data.  
+
+i. 6: add axis label, what type of data is shown?  
+
+Old Fig. 6 was replaced by new Fig. 6 and Fig. 7.
+
+<--- Page Split --->
+
+Reviewer #2 (Remarks to the Author):  
+
+In this manuscript from Vaknin et al, the authors develop a library of synthetic promoters by merging a target core promoter and upstream TF binding sites, to enhance protein expression in eukaryotic cells. This is not the first time synthetic promoters are designed by this approach and previous work has shown that is possible to engineer mammalian cells based on mining of TFBS and core promoter engineering (e.g. Johari et al. 2019 and others).  
+
+The manuscript tries to go beyond the state of art by building a system functional in more than one chassis. I have a few questions, mainly on the experimental side, as I am not a computational person. It is not fully clear to me what the authors mean with the term "motif" and it could be useful to define this.  
+
+We thank the reviewer for this very important comment and apologize for not specifying our rationale for choosing the motifs. In response to the comment, we have added a Methods subsection titled "Motif selection and encoding into OL", where our motif selection process and definition are described. In brief, motifs were based on short conserved sequence segments obtained by Jussi Taipale's group, and provided to us prior to their publication as position weight matrices (DOI: 10.1038/nbt.4138). The Taipale motifs were created by an empirical assay that was designed to test protein activity in a broad swath of organisms.  
+
+It is not fully clear also how the 41 motifs were identified and if this was validated by using more than one TF data base.  
+
+In brief, the Taipale study focused on developing a protein activity assay specifically for DNA- binding TFs in cell and tissue extracts. Based on the study's data, we selected 41 enriched motifs from various organisms and tissue cells (e.g., yeast, fly, and mice tissue cells), according to the following criteria:  
+
+3. Different motif types: 8 organism-shared motifs, 5 mice tissue-shared motifs, 14 unshared motifs (unique to an organism), and 14 unknown motifs (with unknown regulatory function). 
+4. Known/unknown regulatory function: 27 of the selected motifs have known regulatory function, and those were anticipated to be the control motifs. Characterizing the regulatory function of the 14 unknown motifs in yeast was one of the stated goals of the study.  
+
+With regards to validation by an additional database, this was not done for two major reasons. First, we chose the motifs based on the Taipale group experimental findings to show that our approach can be used as a validation exercise to similar experimental findings. Second, 14 of the motifs were uncharacterized and thus would not appear in any database. In fact, our assay enabled us to characterize the regulatory function of 5 of the 14 unknown motifs, and thus validated our approach.  
+
+In figure 1e it is not fully clear why authors have not taken population 2 and re- bin it in order to capture more variants and potentially more interesting candidates with higher expression. It could have been beneficial to do so here.  
+
+We fully agree with the reviewer's sentiment. Unfortunately, in this case, hindsight is 20/20. At the time of the actual SORT- seq experiment, we chose not to re- bin as we were worried about how the increased sorting time may affect the re- binned yeast cells. In addition, in preliminary tests we did not expect to find many variants in this upper bin, and as a result we opted not to risk adding expression noise to the re- binned variants at the expanse of losing expression resolution. Since we were able to identify several strong up- regulating motifs, and train a machine- learning model which provided robust predictions for the unseen validation set, we are confident that whatever information was lost in our decision to not re- bin would not have affected our results or conclusions profoundly.  
+
+In figure 6, for CHO cell experiments, the authors mentioned they used a BFP for fluorescence normalisation. However, it is known from literature that resource competition can impact normalisation (see Frei et al, 2020. Jones et al, 2020).
+
+<--- Page Split --->
+
+We thank the reviewer for this comment and would like to clarify. The use of BFP in the CHO and HeLa cell experiments was done to ensure that cells that were identified as “red” (i.e. mCherry expressing) were indeed transfected by a plasmid and not false positives. In addition, for each variant we compared the mean mCherry expression levels measured to the ratio of the mCherry and BFP channels. The results (see the revised Supplementary Figure 6e) show that BFP does not impact the results or their interpretation.  
+
+Can the authors shown the expression levels of the CHO library but with no normalisation and check if that improves the results?  
+
+We thank the reviewer for this comment. In the revised manuscript, all mammalian data (see new Figs. 6 and 7) is presented without normalization by BFP.  
+
+Caption of figure 6 should also more clearly describe what the figures shows.  
+
+We thank the reviewer for this comment. The captions for revised Fig. 6 and 7 provide additional details about the data presented as compared with the captions in the original manuscript.  
+
+Minor comments pertain to several typos present in the manuscript like Eukaryotic that should be lower case; Figures are called at time with capital letter and at times with lower case.  
+
+We corrected the typos.  
+
+In conclusion I suggest the work to be published once these questions have been addressed.
+
+<--- Page Split --->
+
+Reviewers' Comments:  
+
+Reviewer #1:  
+
+Remarks to the Author:  
+
+The authors addressed all major and minor comments listed in the first round of reviews. The manuscript is greatly improved – I would recommend publication.  
+
+Reviewer #2:  
+
+Remarks to the Author:  
+
+First of all, I would like to thank the authors for addressing my comments and questions. I am generally satisfied with their revision but I would need to ask clarification on two points. Supplementary figure 6. The authors state that they are confirming that the expression of BFP does not impact mCherry expression. However, it is not very clear from the figure caption and response in the rebuttal, nor from the methods section, if they actually performed the experiments in presence and absence of competition or if they simply compared the mCherry signal alone with mCherry normalised on BFP but from the same experiment peformed in presence of competition. Can the authors clarify?  
+
+I would like also to ask the authors if they can comment on the large error bars present for HeLa cell expression experiments.
+
+<--- Page Split --->
+
+## Reviewer #2  
+
+First of all, I would like to thank the authors for addressing my comments and questions. I am generally satisfied with their revision but I would need to ask clarification on two points.  
+
+Reviewer comment: Supplementary figure 6. The authors state that they are confirming that the expression of BFP does not impact mCherry expression. However, it is not very clear from the figure caption and response in the rebuttal, nor from the methods section, if they actually performed the experiments in presence and absence of competition or if they simply compared the mCherry signal alone with mCherry normalised on BFP but from the same experiment performed in presence of competition. Can the authors clarify?  
+
+Author response: We thank the reviewer for the opportunity to clarify. We did not create a separate set of clones lacking the BFP gene. The plot presented in Supplementary Fig. 5e (note the renumbering of the supplementary figures) depicts the mCherry channel compared with the mCherry normalized by the BFP channel as measured on the same cells. The purpose of this plot is to show that normalizing by a synthetic house- keeping gene did not affect the trends observed in the data. We remind the reviewer that the utilization of this house- keeping gene was for us to ensure that sorted cells were transfected by the sURS plasmid. In our opinion an experiment lacking this form of "house- keeping" may not be as reliable, since "false- positive" cells may enter the analysis particularly for weakly expressing strains. We have added a statement regarding this experimental strategy to the Figure caption of supplementary figure 6 to alleviate any further confusion.  
+
+Reviewer comment: I would like also to ask the authors if they can comment on the large error bars present for HeLa cell expression experiments.  
+
+Author response: The error bar observed for the HeLa cells (i.e. Supplementary Fig. 5h) reflect the actual natural deviation in the expression data observed over the triplicates. The unusual variation may be due to the fact that the intensity of expression was not as strong as compared with the expression measured for the CHO cells, which may have led to increased noise. Fortunately, this expression level was sufficiently strong to differentiate between the different boosts enabled by the sURS variants, which correlated well with both the CHO cell measurements and the revised model.
+
+<--- Page Split --->

peer_reviews/12bee38be00ca6148fc0c8a69c26bf2450d3af9d3ae7dfb34f177755640c2e32/supplementary_0_Peer Review File/supplementary_0_Peer Review File_det.mmd

@@ -0,0 +1,623 @@
+<|ref|>title<|/ref|><|det|>[[61, 41, 508, 90]]<|/det|>
+# nature portfolio  
+
+<|ref|>text<|/ref|><|det|>[[70, 111, 362, 140]]<|/det|>
+Peer Review File  
+
+<|ref|>text<|/ref|><|det|>[[70, 155, 918, 211]]<|/det|>
+A universal system for boosting gene expression in Eukaryotic cell- lines  
+
+<|ref|>image<|/ref|><|det|>[[57, 732, 240, 783]]<|/det|>  
+
+<|ref|>text<|/ref|><|det|>[[250, 732, 912, 785]]<|/det|>
+Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work. The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[119, 85, 293, 98]]<|/det|>
+Reviewers' Comments:  
+
+<|ref|>text<|/ref|><|det|>[[119, 112, 223, 125]]<|/det|>
+Reviewer #1:  
+
+<|ref|>text<|/ref|><|det|>[[119, 127, 300, 140]]<|/det|>
+Remarks to the Author:  
+
+<|ref|>text<|/ref|><|det|>[[118, 140, 864, 255]]<|/det|>
+In this study, the authors introduce an algorithm designed to predict regulatory sequences that modulate gene expression in eukaryotic cells. To achieve this, they identified 41 Transcription Factor Binding Sites (TFBS) from various organisms, including S. cerevisiae, S. pombe, D. melanogaster S2 cells, and murine ES cells. These TFBS (or motifs) were used to construct a library of nearly 190,000 variants, each containing up to three different motifs and their variants. Using a combination of machine learning and oligo library analysis in yeast, the authors aimed to identify specific motifs capable of enhancing or attenuating gene expression. These motifs were then assessed in two mammalian cell lines.  
+
+<|ref|>text<|/ref|><|det|>[[119, 254, 833, 308]]<|/det|>
+Overall, the study addresses a topic of general interest: fine tunability of gene expression for optimal output production. However, the manuscript suffers from issues related to clarity, methodology transparency, and validation, ultimately raising concerns about the rigour and robustness of the study. Detailed comments and suggestions are provided below.  
+
+<|ref|>text<|/ref|><|det|>[[118, 321, 876, 405]]<|/det|>
+1. The TFBS used in this study are for endogenous TFs, and some correlate with both activation and repression activity. This may depend on different cell types, cell states and growth conditions, potentially generating high variability in the promoter's behavior. The authors should validate the identified sequences with a significant number of biological replicates (at least 3), at different cell passages and possibly different culturing conditions (e.g. different cell densities, different media composition, etc).  
+
+<|ref|>text<|/ref|><|det|>[[118, 405, 880, 504]]<|/det|>
+2. The authors should ensure that the cells carrying a specific barcode indeed harbour the intended promoter variant, especially considering the complexity of the designed library. Thus, one significant gap in the methodology is the absence of verification steps after sorting, ideally conducted after the final PCR step just before Next-Generation Sequencing (NGS). This validation step could take various forms, even only Sanger sequencing of a select subset of variants (not only 1!). This additional verification would provide confidence in the accuracy of the cell content and the presence of the intended promoter variants.  
+
+<|ref|>text<|/ref|><|det|>[[118, 504, 880, 644]]<|/det|>
+3. Figure 2c should ideally show a relatively tight cluster of data points resembling a straight line with some expected noise and a few outliers, reflecting the inherent variability in biological systems and NGS data. However, the observed data points in this plot appear scattered, indicating a significant divergence from expected behavior. To mitigate this concern and validate the Sort-Seq values, it is advisable to conduct an extensive validation experiment. This experiment could involve the measurement of fluorescence or RNA-seq for individual variants, followed by a correlation analysis with the NGS-based fluorescence (FL mean) values. If this validation does not yield a strong correlation, it could raise doubts about the reliability of the entire experimental approach and data interpretation. Therefore, addressing these issues through thorough verification and validation steps is crucial to enhance the study's credibility and confidence in the results.  
+
+<|ref|>text<|/ref|><|det|>[[118, 644, 864, 685]]<|/det|>
+4. Increased gene expression, can correlate with reduced host organism fitness (DOI: https://doi.org/10.1038/nmeth.3339) and unwanted competition for intracellular resources (DOI: https://doi.org/10.1038/s41467-020-18392-x). If the final goal is to characterize new promoters for bioproduction, the authors should show that they do not impact the fitness of the host cells.  
+
+<|ref|>text<|/ref|><|det|>[[118, 685, 860, 725]]<|/det|>
+5. Both introduction and discussion need more supporting citations. Please add evidences to support all claims. Additionally, some claims seem only partially true or imprecise:  
+
+<|ref|>text<|/ref|><|det|>[[118, 725, 880, 800]]<|/det|>
+a. Line 47: there are studies that addressed the same problem (e.g. DOI: https://doi.org/10.1128/aem.00939-22). These should be cited and discussed. 
+b. Lines 81-85: these claims should be either supported by citations or toned down. There are many well characterized constitutive promoters active across eukaryotic species (CMV, EF1a, PGK, etc). The authors should motivate why their design is superior to existing solutions.  
+
+<|ref|>text<|/ref|><|det|>[[118, 800, 590, 813]]<|/det|>
+6. Some experimental choices and pipeline steps are unclear:  
+
+<|ref|>text<|/ref|><|det|>[[118, 813, 870, 840]]<|/det|>
+a. How were the 41 TFBS selected, especially since many are not common across the 4 eukaryotic cell types evaluated?  
+
+<|ref|>text<|/ref|><|det|>[[118, 840, 632, 854]]<|/det|>
+b. Line 128: how was the threshold of \(>73\%\) determined and why?  
+
+<|ref|>text<|/ref|><|det|>[[118, 854, 875, 895]]<|/det|>
+c. In the method section, describe how the total of 1600 million NGS reads mentioned is processed and the steps involved. Why only \(25\%\) of NGS reads were selected? What is the potential impact on the results?  
+
+<|ref|>text<|/ref|><|det|>[[118, 896, 860, 910]]<|/det|>
+7. Some of the conclusions drawn by the authors in the results section are questionable. Can the
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[120, 84, 441, 98]]<|/det|>
+authors comment on the following points?  
+
+<|ref|>text<|/ref|><|det|>[[116, 98, 872, 530]]<|/det|>
+a. In Fig. 2a the authors claim that there is a clear correlation between the median of the mean fluorescence and the motif: this is not clear from this graphic and it could be pure randomness.  
+b. Line 236: the observed correlation indicates that the distribution is mainly not described by the model.  
+c. Line 269-271 and Fig. 3d: there is no clear correlation between an increase in mixed letters and increase in model performance. The values of 10, 26, 36, 16, 23, 18% of correlation could be more or less random numbers.  
+8. The manuscript presentation is unclear in some parts, especially in the results and discussion sections preventing full understanding of the study:  
+a. The term "mean FL" is a function of the NGS reads. This is confusing as "mean FL" should intuitively be the mean of the measured fluorescence only.  
+b. Is "mean fluorescence" (Fig. 2a) same as "mean FL" (Fig. 2c and others)?  
+c. Large parts of the results section suffer from a lack of clarity, for example lines 168-190. Please work on improving the clarity of the entire section.  
+d. The structure of the discussion is confusing: start with a brief summary of the work, then draw a conclusion and hypothesize the next steps, only to return to the summary in the middle of the second paragraph. Consider restructuring.  
+9. There are several overstatements in the manuscript that should be toned down. E.g.:  
+a. Line 106 "pan-organism": the study focuses on one yeast strain and two mammalian cell lines only.  
+b. Line 112 "different murine tissues": this is inconsistent with the caption of Supplementary Figure 1 where only mouse embryonic stem cells are mentioned.  
+10. Some scientific basics are lacking:  
+a. space between number and units  
+b. organism in italic  
+c. missing descriptions in the methods (see comment 6c)  
+d. missing code and data online: will these be provided after publication?  
+e. the unit of fluorescence is never stated, is it arbitrary units (au)?  
+f. number of biological replicates should be specified. Especially Figure 6 does not have error bars: was validation run on one biological replicate only? If this is the case, it is scientifically inaccurate to draw any conclusion.  
+
+<|ref|>text<|/ref|><|det|>[[118, 545, 252, 558]]<|/det|>
+Minor comments:  
+
+<|ref|>text<|/ref|><|det|>[[115, 559, 868, 884]]<|/det|>
+1. Line 49: "whose promoter is capable of transcribing" is wrong. A promoter does not transcribe, the polymerase does.  
+2. Line 102: "reliable prediction"  
+3. Line 106: "constitutive" instead of non-inducible  
+4. Line 156: "Fig. 1f"  
+5. Line 802: what is a PPM?  
+6. Line 140: Fig. 1c?  
+7. Line 185: which type of correlation is it?  
+8. Line 422: "for a weak and a strong promoter", only one variant tested for each  
+9. Line 430: "circuits" (Remove "bio-")  
+10. Line 869: "of" "s"?  
+11. Lines 780-781: lasers wavelength?  
+12. Discussion: The name UNILIB is introduced only at this point, why?  
+13. Figures:  
+a. Increase the font size in all figures, as they appear too small for easy readability  
+b. 1f: the vertical line should be at 40, it is at 30 though  
+c. 2d: no error bars  
+d. 3e: labels can be put under x axis instead of colors  
+e. 3d: if a linear correlation (Pearson) is assumed, add the line to the plot  
+f. 4a: error bars are at null level?  
+g. 4b (and following): no tick marks in the right plots  
+h. 5f: error bars  
+i. 6: add axis label, what type of data is shown?
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[118, 99, 222, 111]]<|/det|>
+Reviewer #2:  
+
+<|ref|>text<|/ref|><|det|>[[118, 113, 298, 125]]<|/det|>
+Remarks to the Author:  
+
+<|ref|>text<|/ref|><|det|>[[118, 127, 858, 197]]<|/det|>
+In this manuscript from Vaknin et al, the authors develop a library of synthetic promoters by merging a target core promoter and upstream TF binding sites, to enhance protein expression in eukaryotic cells. This is not the first time synthetic promoters are designed by this approach and previous work has shown that is possible to engineer mammalian cells based on mining of TFBS and core promoter engineering (e.g. Johari et al. 2019 and others).  
+
+<|ref|>text<|/ref|><|det|>[[118, 197, 850, 225]]<|/det|>
+The manuscript tries to go beyond the state of art by building a system functional in more than one chassis.  
+
+<|ref|>text<|/ref|><|det|>[[118, 225, 850, 266]]<|/det|>
+I have a few questions, mainly on the experimental side, as I am not a computational person. It is not fully clear to me what the authors mean with the term "motif" and it could be useful to define this.  
+
+<|ref|>text<|/ref|><|det|>[[118, 266, 860, 294]]<|/det|>
+It is not fully clear also how the 41 motifs were identified and if this was validated by using more than one TF data base.  
+
+<|ref|>text<|/ref|><|det|>[[118, 294, 870, 335]]<|/det|>
+In figure 1e it is not fully clear why authors have not taken population 2 and re- bin it in order to capture more variants and potentially more interesting candidates with higher expression. It could have been beneficial to do so here.  
+
+<|ref|>text<|/ref|><|det|>[[118, 335, 844, 377]]<|/det|>
+In figure 6, for CHO cell experiments, the authors mentioned they used a BFP for fluorescence normalisation. However, it is known from literature that resource competition can impact normalisation (see Frei et al, 2020. Jones et al, 2020).  
+
+<|ref|>text<|/ref|><|det|>[[118, 377, 864, 418]]<|/det|>
+Can the authors shown the expression levels of the CHO library but with no normalisation and check if that improves the results? Caption of figure 6 should also more clearly describe what the figures shows.  
+
+<|ref|>text<|/ref|><|det|>[[118, 419, 866, 446]]<|/det|>
+Minor comments pertain to several typos present in the manuscript like Eukaryotic that should be lower case;  
+
+<|ref|>text<|/ref|><|det|>[[118, 447, 680, 461]]<|/det|>
+Figures are called at time with capital letter and at times with lower case.  
+
+<|ref|>text<|/ref|><|det|>[[118, 475, 828, 490]]<|/det|>
+In conclusion I suggest the work to be published once these questions have been addressed.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 90, 392, 106]]<|/det|>
+Reviewer #1 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[114, 128, 882, 305]]<|/det|>
+In this study, the authors introduce an algorithm designed to predict regulatory sequences that modulate gene expression in eukaryotic cells. To achieve this, they identified 41 Transcription Factor Binding Sites (TFBS) from various organisms, including S. cerevisiae, S. pombe, D. melanogaster S2 cells, and murine ES cells. These TFBS (or motifs) were used to construct a library of nearly 190,000 variants, each containing up to three different motifs and their variants. Using a combination of machine learning and oligo library analysis in yeast, the authors aimed to identify specific motifs capable of enhancing or attenuating gene expression. These motifs were then assessed in two mammalian cell lines. Overall, the study addresses a topic of general interest: fine tunability of gene expression for optimal output production. However, the manuscript suffers from issues related to clarity, methodology transparency, and validation, ultimately raising concerns about the rigour and robustness of the study. Detailed comments and suggestions are provided below.  
+
+<|ref|>text<|/ref|><|det|>[[115, 328, 882, 410]]<|/det|>
+1. The TFBS used in this study are for endogenous TFs, and some correlate with both activation and repression activity. This may depend on different cell types, cell states and growth conditions, potentially generating high variability in the promoter's behavior. The authors should validate the identified sequences with a significant number of biological replicates (at least 3), at different cell passages and possibly different culturing conditions (e.g. different cell densities, different media composition, etc).  
+
+<|ref|>text<|/ref|><|det|>[[115, 414, 882, 529]]<|/det|>
+We thank the reviewer for this comment, and the opportunity to improve our work. First, we would like to note that only 27 of the 41 motifs were validated transcription-factor binding sites. The rest of the motifs were unknown, and not attributed to a particular regulatory process. The motifs were found to be conserved across various Eukaryotic lineages in a HT-SELEX assay carried out by Jussi Taipale's group (published in DOI: 10.1038/nbt.4138). As a result of the UNILIB experiment, we were able to characterize 5 of the unknown motifs as either up- or down-regulating. Some of these motifs are weak and require multiple sites or combinations of sites to show a measurable regulatory effect.  
+
+<|ref|>text<|/ref|><|det|>[[115, 534, 882, 631]]<|/det|>
+Second, with regards to the function of the motifs in different cell- growth conditions or cell types, there indeed could be a variation. Consequently, we created 43 unseen synthetic URS variants. We used 32 variants, which sampled 23 of the motifs for which a significant up or down- regulatory behavior was identified, for the results depicted in Fig. 5, and an additional 11 variants were used for the validation of the machine- learning model in Fig. 3. Altogether the 43 validation variants sample 31 of the 42 motifs including all the motifs that were found to be significant.  
+
+<|ref|>text<|/ref|><|det|>[[115, 637, 882, 703]]<|/det|>
+However, given the reviewer's critical comment, we decided to use these variants to expand our validation set to further strengthen our conclusions. The revised manuscript now includes two new figures: Fig. 6 and Fig. 7, which replace the original Fig. 6. New Fig. 6 details the results of 7 separate experimental measurements carried out in biological triplicates as follows:  
+
+<|ref|>text<|/ref|><|det|>[[142, 710, 875, 829]]<|/det|>
+1. Yeast: SD-Ura+2% glucose, \(30^{\circ}\mathrm{C}\) , weak core promoter (mCore), no additives, 43 variants. 
+2. Yeast: SD-Ura+2% glycerol, \(30^{\circ}\mathrm{C}\) , weak core promoter (mCore), no additives, 43 variants. 
+3. Yeast: SD-Ura+2% glucose, \(39^{\circ}\mathrm{C}\) , weak core promoter (mCore), no additives, 43 variants. 
+4. Yeast: SD-Ura+2% glucose, \(30^{\circ}\mathrm{C}\) , weak core promoter (mCore), 1M NaCl, 43 variants. 
+5. Yeast: SD-Ura+2% glucose, \(30^{\circ}\mathrm{C}\) , strong core promoter (FEC-mCore), no additives, 20 variants. 
+6. CHO/K1 cells grown in F12 medium, 32 variants. 
+7. HeLa cells grown in DMEM high glucose medium, 32 variants.  
+
+<|ref|>text<|/ref|><|det|>[[115, 835, 882, 900]]<|/det|>
+The results for all the yeast experiments, whether as biological replicates or in different culturing conditions, correlate strongly with one another in a statistically significant fashion (see below Figs. L1 in addition to Fig. 6 and Supplemental Fig. 6 in the revised manuscript). These results support a conserved function in yeast independent of growth conditions. This is consistent with the previous findings of (Keren et al., Mol.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[114, 90, 882, 219]]<|/det|>
+Sys. Biol., 9:701 (2013)) who tested the activities of \(\sim 900\) S. cerevisiae promoters in 10 different conditions and found no significant difference in activities across the panel of growth conditions used. In addition, the experimental data also shows a statistically significant correlation between the mammalian cell measurements and the yeast measurements, which we showed also in the original manuscript. Finally, we also improved our modelling scheme by extending the machine- learning model developed in Fig. 3 using the additive model described in Fig. 5 to create a hybrid model, which improved performance on all datasets. The new hybrid model now predicts HeLa and CHO cells regulatory activity in a statistically significant fashion (see new Fig. 7).  
+
+<|ref|>text<|/ref|><|det|>[[114, 226, 882, 290]]<|/det|>
+The revised validation segment of our manuscript now allows us to make an improved argument that the synthetic URS sequences that we created can generate a regulatory function that is, for the most- part, independent of cell- type and growth conditions, making these sequences a potentially widely applicable tool.  
+
+<|ref|>image<|/ref|><|det|>[[123, 308, 876, 504]]<|/det|>
+<|ref|>image_caption<|/ref|><|det|>[[114, 533, 844, 566]]<|/det|>
+<center>Figure L1: Comparison of biological triplicates for the 43 validation variants grown in SD-\(Ura + 2\%\) glucose. Experiments show a strong a correlation between all 3 repeats. </center>  
+
+<|ref|>text<|/ref|><|det|>[[114, 580, 882, 677]]<|/det|>
+2. The authors should ensure that the cells carrying a specific barcode indeed harbour the intended promoter variant, especially considering the complexity of the designed library. Thus, one significant gap in the methodology is the absence of verification steps after sorting, ideally conducted after the final PCR step just before Next-Generation Sequencing (NGS). This validation step could take various forms, even only Sanger sequencing of a select subset of variants (not only 1!). This additional verification would provide confidence in the accuracy of the cell content and the presence of the intended promoter variants.  
+
+<|ref|>text<|/ref|><|det|>[[114, 692, 882, 853]]<|/det|>
+In brief, due to their short length (bc+sURS+core- promoter) our variants were fully sequenced in the NGS. While we apologize for not including this data in the manuscript, as part of our validation, we screened for all full- length reads in the library (i.e. reads that yielded a full length of over 200 bp that include the barcodes, sURS, and primer regions). We found approx. 400M such reads, and they all perfectly matched the barcodes, motifs, and primer regions. We then quantified the correctness of the downstream sequence to the barcode with the intended design and found that per position \(>99\%\) match the design. In addition, we designed the barcodes, such that the Hamming distance between each two is at least 3, which maximized our ability to identify variants. We thank the reviewer for this important comment. In response to the comment, we have added a comment under the Methods sub- section "NGS data processing and read normalization" to clarify this issue.  
+
+<|ref|>text<|/ref|><|det|>[[113, 869, 881, 902]]<|/det|>
+3. Figure 2c should ideally show a relatively tight cluster of data points resembling a straight line with some expected noise and a few outliers, reflecting the inherent variability in biological systems and NGS data.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 90, 882, 202]]<|/det|>
+However, the observed data points in this plot appear scattered, indicating a significant divergence from expected behavior. To mitigate this concern and validate the Sort- Seq values, it is advisable to conduct an extensive validation experiment. This experiment could involve the measurement of fluorescence or RNA- seq for individual variants, followed by a correlation analysis with the NGS- based fluorescence (FL mean) values. If this validation does not yield a strong correlation, it could raise doubts about the reliability of the entire experimental approach and data interpretation. Therefore, addressing these issues through thorough verification and validation steps is crucial to enhance the study's credibility and confidence in the results.  
+
+<|ref|>text<|/ref|><|det|>[[115, 209, 882, 272]]<|/det|>
+We are aware of the lack of correlation observed in our OL measurements, which we view as a feature of the methodology rather than an artefact. The data used to compose panel 2c consisted of OL variants that were encoded with two barcodes. The lack of correlation observed for most variant pairs was due to three main reasons:  
+
+<|ref|>text<|/ref|><|det|>[[144, 273, 883, 449]]<|/det|>
+1. In SORT-seq experiments, it is impossible to tell how many reads came from a single sorted cell. For example, one cell can yield several hundred reads or a single read in the final count. This can lead to a lack of correlation if the number of cells sorted per a particular variant is small as in this case. We remind the reviewer that our library was unprecedented in size ( \(\sim 200\mathrm{K}\) mixed-base variants), and thus the number of cells collected for each variant was small. 
+2. In SORT-seq experiments, it is impossible to tell whether cells die after sorting or divide. In such a case, a single cell may also yield more read events in comparison to another cell, though the original sort had them at an equal weight. 
+3. Variants which do not encode a particular regulatory function are more susceptible to transcriptional variation. The strongly up- and down-regulating variants (top and bottom corners of the plot respectively) are strongly correlated as expected from a fully functional regulatory element.  
+
+<|ref|>text<|/ref|><|det|>[[115, 464, 882, 576]]<|/det|>
+The analysis shown in Figure 2 and referred to by the reviewer was, therefore, used as impetus for the quantitative analysis that was done on the 2,435- variant subset that were encoded with 22 barcodes. We reasoned that the sources for the lack of correlation discussed in the points above will be minimized by analyzing multiple barcodes on the same variants. For example, identifying all 22 barcodes immediately implies that at the very least we screened 22 separate fluorescent yeast cells. Consequently, we reasoned that such a subset will lead to a more robust statistical analysis, even though a given pair of barcodes may not be as correlated as we would like.  
+
+<|ref|>text<|/ref|><|det|>[[115, 592, 882, 704]]<|/det|>
+The analysis carried out in Figure 3 shows that the choice of unifying the 22 barcodes is statistically robust as evident by the agreement with the de Boer machine- learning model that was trained using independent experimental data. In addition, the validation set of Figures 5- 7 and the agreement with the models that were derived from the experimental data obtained from the 2,435- variant subset essentially shows that both the SORT- seq experiment and the analysis were robust. Consequently, we believe that we have provided sufficient proof in the manuscript both computationally and experimentally that our analysis and experimental approach were robust, and no additional validation is necessary.  
+
+<|ref|>text<|/ref|><|det|>[[115, 720, 882, 785]]<|/det|>
+4. Increased gene expression, can correlate with reduced host organism fitness (DOI: https://doi.org/10.1038/nmeth.3339) and unwanted competition for intracellular resources (DOI: https://doi.org/10.1038/s41467-020-18392-x). If the final goal is to characterize new promoters for bioproduction, the authors should show that they do not impact the fitness of the host cells.  
+
+<|ref|>text<|/ref|><|det|>[[115, 791, 882, 888]]<|/det|>
+We thank the reviewer for this comment. In response to the comment, we carried out fitness analysis on the 32 validation variants. The results are shown in Figure L2 and have been added to the Supplementary Information as Supplementary Fig. 6. In our experiments, we did not detect any effect on fitness as a result of our synthetic upstream regulatory sequences and mentioned that in the "Boosted expression level translates from yeast and mammalian cells" Results subsection of the revised manuscript. Specifically, we grew the 43 validation variants in YPD and tracked their growth via OD measurement as function of time
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[113, 90, 883, 123]]<|/det|>
+(Fig. L2a - circles). For each strain we then fitted the OD measurements (Fig. L2a - blue line) with the following model for exponential growth:  
+
+<|ref|>equation<|/ref|><|det|>[[374, 123, 621, 160]]<|/det|>
+\[OD_{600}(t) = C + \frac{L}{(1 + e^{-k(t - t_0)})}\]  
+
+<|ref|>text<|/ref|><|det|>[[114, 172, 883, 285]]<|/det|>
+Where \(C\) is background OD levels, \(L\) is the max OD, \(k\) is the growth rate, and \(t_0\) corresponds to the lag time (i.e. time at which the culture reaches OD of L/2). Using this model, we extracted the growth rate for each strain and for both repeats. We plotted in Fig. L2b the different fitted growth rates \((k)\) for both repeats that were measured for each variant. The results show that the growth rates for all experiments was found to be within a narrow range of \(\sim 0.4 - 1.3\) (1/hr) without any significant correlation between duplicates. The lack of correlation between duplicates and narrow range of growth rates provides strong evidence that our variants do not affect the fitness of the yeast cells.  
+
+<|ref|>image<|/ref|><|det|>[[123, 298, 861, 581]]<|/det|>
+<|ref|>image_caption<|/ref|><|det|>[[114, 584, 883, 649]]<|/det|>
+<center>Figure L2: Fitness analysis for yeast cells expressing the 43-variant validation set grown in YPD. (a) The 43 validation variants were grown in YPD in duplicates and tracked for OD as a function of time (circles). For each variant the growth data are fitted (blue lines) by a classic growth curve (see Supplementary Information). (b) The rates of growth for each variant are plotted as a scatter plot pair for both repeats. </center>  
+
+<|ref|>text<|/ref|><|det|>[[113, 662, 865, 775]]<|/det|>
+5. Both introduction and discussion need more supporting citations. Please add evidences to support all claims. Additionally, some claims seem only partially true or imprecise: a. Line 47: there are studies that addressed the same problem (e.g. DOI: https://doi.org/10.1128/aem.00939-22). These should be cited and discussed. b. Lines 81-85: these claims should be either supported by citations or toned down. There are many well characterized constitutive promoters active across eukaryotic species (CMV, EF1a, PGK, etc). The authors should motivate why their design is superior to existing solutions.  
+
+<|ref|>text<|/ref|><|det|>[[115, 790, 883, 855]]<|/det|>
+We thank the reviewer for providing the additional reference, and for pointing out misstatements in the Introduction. In response to the comment, we added this and other references and rephrased the text in the Introduction and Discussion sections accordingly. Please see the marked up version of the revised manuscript for the details.  
+
+<|ref|>text<|/ref|><|det|>[[114, 871, 555, 888]]<|/det|>
+6. Some experimental choices and pipeline steps are unclear:
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 90, 881, 123]]<|/det|>
+a. How were the 41 TFBS selected, especially since many are not common across the 4 eukaryotic cell types evaluated?  
+
+<|ref|>text<|/ref|><|det|>[[115, 129, 882, 305]]<|/det|>
+We apologize that these choices were unclear. In brief, the motifs were chosen as part of a past collaboration with Jussi Taipale's group. At the time, the Taipale group had a study (later published in DOI: 10.1038/nbt.4138) that was focused on developing a protein activity assay specifically for DNA- binding TFs in cell and tissue extracts. His team was able to identify strong and enriched TFBS motifs using this assay, including the 41 TFBSs that we selected for our research. Those and other motifs were discovered in numerous organisms, including two types of yeast, different tissues from mice, bacteria, and Drosophila S2 cells. Even though some of the motifs were not found to be common/conserved across these organisms, we nevertheless chose to characterize them due to similarities to well- characterized protein families in higher eukaryotes (e.g. the bHLH motif). For a detailed explanation as to how and why we chose these 41 motifs, please see the new Methods subsection titled "Motif selection and encoding into OL. Based on the study's data, we selected the 41 enriched motifs according to the following criteria:  
+
+<|ref|>text<|/ref|><|det|>[[144, 311, 882, 393]]<|/det|>
+1. Different motif types: 8 organism-shared motifs, 5 mice tissue-shared motifs, 14 unshared motifs (unique to an organism), and 14 unknown motifs (with unknown regulatory function). 
+2. Known/unknown regulatory function: 27 of the selected motifs have known regulatory function, and those were anticipated to be the control motifs. Characterizing the regulatory function of the 14 unknown motifs in yeast was one of the stated goals of the study.  
+
+<|ref|>text<|/ref|><|det|>[[115, 399, 596, 416]]<|/det|>
+b. Line 128: how was the threshold of \(>73\%\) determined and why?  
+
+<|ref|>text<|/ref|><|det|>[[115, 422, 883, 552]]<|/det|>
+We thank the reviewer for this question and apologize for not including the rationale for our choice in the original manuscript. In response to the comment, we have added a new Methods subsection entitled "Motif selection and encoding into OL", where the rationale is now described. In brief, K and M substitutions were based on the percentage calculation of the respective G/T and A/C occurrences, in each position of the motif according to the PFM data given to us by the Taipale group (later published in DOI: 10.1038/nbt.4138). Positions within the motif, with dominant percentages, were replaced by either K or M in the final design. \(70\%\) threshold was set to determine the K/M substitution, but the actual calculated threshold was higher at \(73\%\) , as specified in the Results section.  
+
+<|ref|>text<|/ref|><|det|>[[115, 566, 882, 599]]<|/det|>
+c. In the method section, describe how the total of 1600 million NGS reads mentioned is processed and the steps involved. Why only \(25\%\) of NGS reads were selected? What is the potential impact on the results?  
+
+<|ref|>text<|/ref|><|det|>[[115, 606, 882, 750]]<|/det|>
+We thank the reviewer for this question. In response to the comment, we have expanded the Methods subsection titled "NGS data processing and read normalization", specifying the processing steps accordingly. Given the size of the UNILIB library, we opted to use Illumina's NovaSeq for its ability to generate as much as 1B reads per run. We used Illumina's S1 kit, and despite Illumina's reliability claims, indeed only \(25\%\) of the reads were found to be correct in both the NovaSeq runs that were made. Neither we nor the Genomics Center at the Weizmann Institute are certain about the underlying cause for this low fidelity, which may have been due to the choice of primers. Irrespective of that, the center offered to re- sequence the library, ultimately allowing us to extract 400M correct reads. This number of reads proved to be sufficient for the analysis in our study, which was based on the machine- learning modelling.  
+
+<|ref|>text<|/ref|><|det|>[[115, 765, 882, 895]]<|/det|>
+As discussed above, our analysis showed that the use of two barcodes per variant was insufficient for proper quantitative analysis, as such a low number introduced a large statistical uncertainty regarding the actual number of sorted cells. On the flip side, the same analysis showed that 22 barcodes per variant significantly reduced this uncertainty. In the former case, increasing the number of reads would not have affected the cell number uncertainty, while in the latter the 22 barcodes ensured that we have a sufficient amount of reads for every one of the variants of the 2,435- variant subset even with only 400M correct reads extracted in total. Consequently, the \(25\%\) extraction rate did not impact our conclusions or results of the downstream analysis.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 105, 883, 139]]<|/det|>
+7. Some of the conclusions drawn by the authors in the results section are questionable. Can the authors comment on the following points?  
+
+<|ref|>text<|/ref|><|det|>[[115, 145, 883, 179]]<|/det|>
+a. In Fig. 2a the authors claim that there is a clear correlation between the median of the mean fluorescence and the motif: this is not clear from this graphic and it could be pure randomness.  
+
+<|ref|>text<|/ref|><|det|>[[115, 185, 882, 329]]<|/det|>
+We thank the reviewer for this comment and apologize for not assessing the correlation in the original manuscript. In response to the comment, to assess the correlation between the median of mean fluorescent expression levels and motifs, we conducted a Wilcoxon rank-sum test by comparing the mean fluorescent expression levels of the group of variants containing each motif to the group of variants containing motifs ranked at least 5 motifs away. To correct for multiple tests, we applied the Benjamini-Hochberg procedure with an FDR threshold of 0.1. This new analysis revealed significant differences for 36 out of 42 motifs (p- value<0.05), where the 6 non- significant motifs are in motifs ordered 15- 20 in decreasing order, which is expected since their values are distributed around the center of the mean- fluorescence distribution. We added this new statistical significance analysis to the Results section.  
+
+<|ref|>text<|/ref|><|det|>[[115, 344, 870, 361]]<|/det|>
+b. Line 236: the observed correlation indicates that the distribution is mainly not described by the model.  
+
+<|ref|>text<|/ref|><|det|>[[115, 367, 882, 544]]<|/det|>
+We thank the reviewer for this comment. In various biological problems, such as inference of protein- DNA binding preferences based on protein- binding- microarray data, correlation values are within 0.4- 0.6 (10.1109/TCBB.2019.2947461). Moreover, the correlation of 0.45 was achieved by the all- data model (ADM), which we showed is inferior to the AMM and MBO models. In addition, this value is calculated on a test set of variants, where most have only 2 barcodes, which implies the lower quality of their mean- fluorescence measurements compared to variants with 22 barcodes. One of the key insights from our study was that variants supported by 22 barcodes, as opposed to 2 barcodes, yielded more accurate mean- fluorescence measurements (due to the inclusion of more cells carrying the variant). When the MBO model was exclusively trained and tested on datasets of variants with 22 barcodes, the obtained Pearson correlation was 0.61. This suggests that using more accurate training and test sets results in a higher observed correlation, which better explains the variability in the data.  
+
+<|ref|>text<|/ref|><|det|>[[115, 559, 882, 608]]<|/det|>
+c. Line 269-271 and Fig. 3d: there is no clear correlation between an increase in mixed letters and increase in model performance. The values of 10, 26, 36, 16, 23, 18% of correlation could be more or less random numbers.  
+
+<|ref|>text<|/ref|><|det|>[[115, 614, 882, 663]]<|/det|>
+We thank the reviewer for pointing this unclear part of our manuscript. We acknowledge that the reported correlation values may not achieve statistical significance. In response to this comment, we removed Fig. 3d and the corresponding text describing the analyses and results from the Results section.  
+
+<|ref|>text<|/ref|><|det|>[[115, 679, 882, 712]]<|/det|>
+8. The manuscript presentation is unclear in some parts, especially in the results and discussion sections preventing full understanding of the study:  
+
+<|ref|>text<|/ref|><|det|>[[115, 712, 882, 744]]<|/det|>
+a. The term "mean FL" is a function of the NGS reads. This is confusing as "mean FL" should intuitively be the mean of the measured fluorescence only.  
+
+<|ref|>text<|/ref|><|det|>[[115, 750, 882, 815]]<|/det|>
+We thank the reviewer for this comment and have changed all "mean FL" labels for the NGS data to the one used by (Sharon, E. et al. Nat Biotechnol 30, 521-530 (2012)) in a similar oligo library experiment: "Expression (A.U.)". All relevant labels in the figures and text have been changed to Expression (A.U.) in the revised manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[115, 831, 736, 864]]<|/det|>
+b. Is "mean fluorescence" (Fig. 2a) same as "mean FL" (Fig. 2c and others)? We thank the reviewer for finding this minor error, and changed to Expression (A.U.).
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[112, 105, 870, 139]]<|/det|>
+c. Large parts of the results section suffer from a lack of clarity, for example lines 168-190. Please work on improving the clarity of the entire section.  
+
+<|ref|>text<|/ref|><|det|>[[115, 138, 878, 187]]<|/det|>
+We thank the reviewer for this comment and have clarified the Results, including the subsection titled "Variants manifest a broad range of regulatory behavior". Please see the marked-up version of the revised manuscript for the changes.  
+
+<|ref|>text<|/ref|><|det|>[[115, 201, 882, 251]]<|/det|>
+d. The structure of the discussion is confusing: start with a brief summary of the work, then draw a conclusion and hypothesize the next steps, only to return to the summary in the middle of the second paragraph. Consider restructuring.  
+
+<|ref|>text<|/ref|><|det|>[[115, 256, 882, 306]]<|/det|>
+We thank the reviewer for this comment. We decided to leave the structure of the Discussion as is with minor modifications. We opted for this option so we can highlight at the beginning of the section the design algorithm that was developed, which we believe is the most important achievement of this work.  
+
+<|ref|>text<|/ref|><|det|>[[115, 320, 882, 355]]<|/det|>
+9. There are several overstatements in the manuscript that should be toned down. E.g.: a. Line 106 "pan-organism": the study focuses on one yeast strain and two mammalian cell lines only.  
+
+<|ref|>text<|/ref|><|det|>[[115, 359, 732, 377]]<|/det|>
+In response to the comment, we deleted the term pan-organism from the Introduction.  
+
+<|ref|>text<|/ref|><|det|>[[115, 391, 882, 425]]<|/det|>
+b. Line 112 "different murine tissues": this is inconsistent with the caption of Supplementary Figure 1 where only mouse embryonic stem cells are mentioned.  
+
+<|ref|>text<|/ref|><|det|>[[115, 431, 882, 465]]<|/det|>
+In response to the comment, we corrected the error and replaced the term "different murine tissues" by mouse (ES cells and different tissues).  
+
+<|ref|>text<|/ref|><|det|>[[116, 480, 393, 497]]<|/det|>
+10. Some scientific basics are lacking:  
+
+<|ref|>text<|/ref|><|det|>[[115, 504, 370, 520]]<|/det|>
+a. space between number and units  
+
+<|ref|>text<|/ref|><|det|>[[115, 528, 277, 543]]<|/det|>
+Errors were corrected.  
+
+<|ref|>text<|/ref|><|det|>[[115, 551, 262, 567]]<|/det|>
+b. organism in italic  
+
+<|ref|>text<|/ref|><|det|>[[115, 575, 277, 590]]<|/det|>
+Errors were corrected.  
+
+<|ref|>text<|/ref|><|det|>[[115, 597, 525, 614]]<|/det|>
+c. missing descriptions in the methods (see comment 6c)  
+
+<|ref|>text<|/ref|><|det|>[[115, 620, 561, 637]]<|/det|>
+All missing methods were added. Please see responses above.  
+
+<|ref|>text<|/ref|><|det|>[[115, 644, 644, 661]]<|/det|>
+d. missing code and data online: will these be provided after publication?  
+
+<|ref|>text<|/ref|><|det|>[[115, 667, 880, 685]]<|/det|>
+All data and code were uploaded. Please see data and code availability statements in the revised manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[115, 691, 595, 708]]<|/det|>
+e. the unit of fluorescence is never stated, is it arbitrary units (au)?  
+
+<|ref|>text<|/ref|><|det|>[[115, 714, 882, 747]]<|/det|>
+The mean FL was replaced by the consensus Expression (A.U.) for all mean level expression level measurements.  
+
+<|ref|>text<|/ref|><|det|>[[115, 753, 882, 803]]<|/det|>
+f. number of biological replicates should be specified. Especially Figure 6 does not have error bars: was validation run on one biological replicate only? If this is the case, it is scientifically inaccurate to draw any conclusion.  
+
+<|ref|>text<|/ref|><|det|>[[115, 809, 882, 906]]<|/det|>
+We thank the reviewer for this comment and would like to clarify our choice of plotting. All validation experiments were conducted in (at least) biological triplicates as discussed above (see Fig. L1 for example). In response to Comment 1, we have substantially increased the volume of the validation experiments, and consequently split the original Fig. 6 in the original manuscript to a new Fig. 6 depicting only experimental data, and a new Figure 7 depicting the modelling analysis. Given the quantity of data, we opted to omit the error-bars from some of the plots for aesthetic reasons (see for example Fig. L3 which depicts Fig. 6a-c
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 90, 883, 123]]<|/det|>
+with error bars). Since we recognize the need to visually and numerically report the variance in the validation measurements, we have done the following:  
+
+<|ref|>text<|/ref|><|det|>[[142, 130, 884, 373]]<|/det|>
+- The variation for the various validation datasets across the repeats can be assessed using new Supplementary Figure 6, and via the correlation heatmap plots of new Fig. 6 (panels d and e). Note, that the autocorrelation between the various repeats of the yeast experiments does not yield a higher Pearson correlation coefficient, as compared with the cross-correlation of the mean fluorescent expression levels of the various conditions. This result provides statistically significant experimental evidence that the different yeast growth conditions do not alter the expression of our variants, and in our opinion is depicted more convincingly via the mode of plotting that we chose to use in the revised manuscript.- We plotted Fig.6a with error bars to convey directly the variance in the cross-correlation analysis. This panel is identical to the one shown in Fig. L3a.- Fig. 6b-c are depicted without error-bars. Here, the purpose is to highlight the fact that out of the dataset 24 variants behaved similarly to yeast (blue), while 8 were uncorrelated in CHO (red). We have added the panel b and c to supplementary fig.6 as panel g and h.- We plotted the error-bars for the both the yeast-glucose \(2\%\) and HeLa-blue variants in panel b and c of Fig. 7.  
+
+<|ref|>image<|/ref|><|det|>[[123, 383, 863, 607]]<|/det|>
+<|ref|>image_caption<|/ref|><|det|>[[114, 621, 883, 703]]<|/det|>
+<center>Figure L3:Cross correlation data for 43-variant validation experiments plotted with error-bars. (a) yeast cross-correlation data. (b-c) The \(2\%\) glucose yeast expression data plotted as a function of CHO (b) and HeLa (c) expression data. Error-bars were computed using standard-error analysis carried out on mean flow cytometry fluorescence measurements obtained from three or four biological repeats – depending on data set. </center>  
+
+<|ref|>text<|/ref|><|det|>[[115, 727, 245, 742]]<|/det|>
+Minor comments:  
+
+<|ref|>text<|/ref|><|det|>[[112, 749, 884, 784]]<|/det|>
+1. Line 49: "whose promoter is capable of transcribing" is wrong. A promoter does not transcribe, the polymerase does.  
+
+<|ref|>text<|/ref|><|det|>[[115, 789, 720, 806]]<|/det|>
+We changed the statement to "whose promoter is capable of initiating transcription"  
+
+<|ref|>text<|/ref|><|det|>[[115, 806, 355, 821]]<|/det|>
+2. Line 102: "reliable prediction"  
+
+<|ref|>text<|/ref|><|det|>[[115, 822, 356, 836]]<|/det|>
+We removed the word prediction.  
+
+<|ref|>text<|/ref|><|det|>[[115, 837, 489, 852]]<|/det|>
+3. Line 106: "constitutive" instead of non-inducible  
+
+<|ref|>text<|/ref|><|det|>[[115, 853, 384, 868]]<|/det|>
+We retrained the word non-inducible.  
+
+<|ref|>text<|/ref|><|det|>[[115, 870, 273, 884]]<|/det|>
+4. Line 156: "Fig. 1f"
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 91, 279, 106]]<|/det|>
+We corrected the typo.  
+
+<|ref|>text<|/ref|><|det|>[[115, 107, 322, 122]]<|/det|>
+5. Line 802: what is a PPM?  
+
+<|ref|>text<|/ref|><|det|>[[115, 123, 520, 139]]<|/det|>
+We corrected PPM to PWM - position weighted matrix.  
+
+<|ref|>text<|/ref|><|det|>[[115, 140, 272, 155]]<|/det|>
+6. Line 140: Fig. 1c?  
+
+<|ref|>text<|/ref|><|det|>[[115, 156, 279, 170]]<|/det|>
+We corrected the typo.  
+
+<|ref|>text<|/ref|><|det|>[[115, 171, 430, 186]]<|/det|>
+7. Line 185: which type of correlation is it?  
+
+<|ref|>text<|/ref|><|det|>[[115, 187, 339, 201]]<|/det|>
+We added the word "Pearson".  
+
+<|ref|>text<|/ref|><|det|>[[115, 202, 700, 218]]<|/det|>
+8. Line 422: "for a weak and a strong promoter", only one variant tested for each  
+
+<|ref|>text<|/ref|><|det|>[[115, 219, 615, 234]]<|/det|>
+We changed the phrase in accordance with the reviewer's suggestion.  
+
+<|ref|>text<|/ref|><|det|>[[115, 235, 409, 250]]<|/det|>
+9. Line 430: "circuits" (Remove "bio- ")  
+
+<|ref|>text<|/ref|><|det|>[[115, 251, 304, 265]]<|/det|>
+We removed the word bio.  
+
+<|ref|>text<|/ref|><|det|>[[115, 266, 293, 281]]<|/det|>
+10. Line 869: "of""s"?!  
+
+<|ref|>text<|/ref|><|det|>[[115, 283, 279, 297]]<|/det|>
+We corrected the typo.  
+
+<|ref|>text<|/ref|><|det|>[[115, 298, 395, 313]]<|/det|>
+11. Lines 780-781: lasers wavelength?  
+
+<|ref|>text<|/ref|><|det|>[[115, 315, 405, 329]]<|/det|>
+We added MacsQuant laser wavelength.  
+
+<|ref|>text<|/ref|><|det|>[[115, 330, 642, 346]]<|/det|>
+12. Discussion: The name UNILIB is introduced only at this point, why?  
+
+<|ref|>text<|/ref|><|det|>[[115, 346, 870, 377]]<|/det|>
+We chose to name the completed design algorithm UNILIB, and name it only after we demonstrated that it worked via the validation experiments.  
+
+<|ref|>text<|/ref|><|det|>[[115, 378, 201, 393]]<|/det|>
+13. Figures:  
+
+<|ref|>text<|/ref|><|det|>[[115, 394, 707, 410]]<|/det|>
+a. Increase the font size in all figures, as they appear too small for easy readability  
+
+<|ref|>text<|/ref|><|det|>[[115, 411, 325, 425]]<|/det|>
+We corrected where possible.  
+
+<|ref|>text<|/ref|><|det|>[[115, 426, 515, 441]]<|/det|>
+b. 1f: the vertical line should be at 40, it is at 30 though  
+
+<|ref|>text<|/ref|><|det|>[[115, 442, 280, 456]]<|/det|>
+We corrected this error.  
+
+<|ref|>text<|/ref|><|det|>[[115, 458, 253, 472]]<|/det|>
+c. 2d: no error bars  
+
+<|ref|>text<|/ref|><|det|>[[115, 473, 790, 489]]<|/det|>
+There are no error bars for the medians of the mean fluorescent expression level distributions.  
+
+<|ref|>text<|/ref|><|det|>[[115, 490, 500, 505]]<|/det|>
+d. 3e: labels can be put under x axis instead of colors  
+
+<|ref|>text<|/ref|><|det|>[[115, 506, 380, 521]]<|/det|>
+We improved the figure accordingly.  
+
+<|ref|>text<|/ref|><|det|>[[115, 522, 640, 538]]<|/det|>
+e. 3d: if a linear correlation (Pearson) is assumed, add the line to the plot  
+
+<|ref|>text<|/ref|><|det|>[[115, 539, 295, 553]]<|/det|>
+We removed old Fig. 3d.  
+
+<|ref|>text<|/ref|><|det|>[[115, 554, 352, 569]]<|/det|>
+f. 4a: error bars are at null level?  
+
+<|ref|>text<|/ref|><|det|>[[115, 570, 863, 600]]<|/det|>
+There are no error bars for the large colored bars as they represent the p- value estimate from comparing two distributions as shown in the inset.  
+
+<|ref|>text<|/ref|><|det|>[[115, 601, 504, 616]]<|/det|>
+g. 4b (and following): no tick marks in the right plots  
+
+<|ref|>text<|/ref|><|det|>[[115, 617, 285, 631]]<|/det|>
+We corrected this error.  
+
+<|ref|>text<|/ref|><|det|>[[115, 633, 230, 647]]<|/det|>
+h. 5f: error bars  
+
+<|ref|>text<|/ref|><|det|>[[115, 648, 870, 679]]<|/det|>
+We opted to present the actual measurements as circles instead of error bars to convey the distribution of the measured data.  
+
+<|ref|>text<|/ref|><|det|>[[115, 680, 459, 696]]<|/det|>
+i. 6: add axis label, what type of data is shown?  
+
+<|ref|>text<|/ref|><|det|>[[115, 697, 472, 712]]<|/det|>
+Old Fig. 6 was replaced by new Fig. 6 and Fig. 7.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 90, 392, 107]]<|/det|>
+Reviewer #2 (Remarks to the Author):  
+
+<|ref|>text<|/ref|><|det|>[[115, 113, 882, 194]]<|/det|>
+In this manuscript from Vaknin et al, the authors develop a library of synthetic promoters by merging a target core promoter and upstream TF binding sites, to enhance protein expression in eukaryotic cells. This is not the first time synthetic promoters are designed by this approach and previous work has shown that is possible to engineer mammalian cells based on mining of TFBS and core promoter engineering (e.g. Johari et al. 2019 and others).  
+
+<|ref|>text<|/ref|><|det|>[[115, 200, 882, 249]]<|/det|>
+The manuscript tries to go beyond the state of art by building a system functional in more than one chassis. I have a few questions, mainly on the experimental side, as I am not a computational person. It is not fully clear to me what the authors mean with the term "motif" and it could be useful to define this.  
+
+<|ref|>text<|/ref|><|det|>[[115, 255, 882, 353]]<|/det|>
+We thank the reviewer for this very important comment and apologize for not specifying our rationale for choosing the motifs. In response to the comment, we have added a Methods subsection titled "Motif selection and encoding into OL", where our motif selection process and definition are described. In brief, motifs were based on short conserved sequence segments obtained by Jussi Taipale's group, and provided to us prior to their publication as position weight matrices (DOI: 10.1038/nbt.4138). The Taipale motifs were created by an empirical assay that was designed to test protein activity in a broad swath of organisms.  
+
+<|ref|>text<|/ref|><|det|>[[115, 359, 882, 391]]<|/det|>
+It is not fully clear also how the 41 motifs were identified and if this was validated by using more than one TF data base.  
+
+<|ref|>text<|/ref|><|det|>[[115, 398, 882, 447]]<|/det|>
+In brief, the Taipale study focused on developing a protein activity assay specifically for DNA- binding TFs in cell and tissue extracts. Based on the study's data, we selected 41 enriched motifs from various organisms and tissue cells (e.g., yeast, fly, and mice tissue cells), according to the following criteria:  
+
+<|ref|>text<|/ref|><|det|>[[143, 454, 882, 536]]<|/det|>
+3. Different motif types: 8 organism-shared motifs, 5 mice tissue-shared motifs, 14 unshared motifs (unique to an organism), and 14 unknown motifs (with unknown regulatory function). 
+4. Known/unknown regulatory function: 27 of the selected motifs have known regulatory function, and those were anticipated to be the control motifs. Characterizing the regulatory function of the 14 unknown motifs in yeast was one of the stated goals of the study.  
+
+<|ref|>text<|/ref|><|det|>[[115, 541, 882, 621]]<|/det|>
+With regards to validation by an additional database, this was not done for two major reasons. First, we chose the motifs based on the Taipale group experimental findings to show that our approach can be used as a validation exercise to similar experimental findings. Second, 14 of the motifs were uncharacterized and thus would not appear in any database. In fact, our assay enabled us to characterize the regulatory function of 5 of the 14 unknown motifs, and thus validated our approach.  
+
+<|ref|>text<|/ref|><|det|>[[115, 628, 882, 677]]<|/det|>
+In figure 1e it is not fully clear why authors have not taken population 2 and re- bin it in order to capture more variants and potentially more interesting candidates with higher expression. It could have been beneficial to do so here.  
+
+<|ref|>text<|/ref|><|det|>[[115, 684, 882, 813]]<|/det|>
+We fully agree with the reviewer's sentiment. Unfortunately, in this case, hindsight is 20/20. At the time of the actual SORT- seq experiment, we chose not to re- bin as we were worried about how the increased sorting time may affect the re- binned yeast cells. In addition, in preliminary tests we did not expect to find many variants in this upper bin, and as a result we opted not to risk adding expression noise to the re- binned variants at the expanse of losing expression resolution. Since we were able to identify several strong up- regulating motifs, and train a machine- learning model which provided robust predictions for the unseen validation set, we are confident that whatever information was lost in our decision to not re- bin would not have affected our results or conclusions profoundly.  
+
+<|ref|>text<|/ref|><|det|>[[115, 828, 861, 877]]<|/det|>
+In figure 6, for CHO cell experiments, the authors mentioned they used a BFP for fluorescence normalisation. However, it is known from literature that resource competition can impact normalisation (see Frei et al, 2020. Jones et al, 2020).
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[115, 90, 882, 170]]<|/det|>
+We thank the reviewer for this comment and would like to clarify. The use of BFP in the CHO and HeLa cell experiments was done to ensure that cells that were identified as “red” (i.e. mCherry expressing) were indeed transfected by a plasmid and not false positives. In addition, for each variant we compared the mean mCherry expression levels measured to the ratio of the mCherry and BFP channels. The results (see the revised Supplementary Figure 6e) show that BFP does not impact the results or their interpretation.  
+
+<|ref|>text<|/ref|><|det|>[[115, 185, 882, 218]]<|/det|>
+Can the authors shown the expression levels of the CHO library but with no normalisation and check if that improves the results?  
+
+<|ref|>text<|/ref|><|det|>[[115, 225, 881, 258]]<|/det|>
+We thank the reviewer for this comment. In the revised manuscript, all mammalian data (see new Figs. 6 and 7) is presented without normalization by BFP.  
+
+<|ref|>text<|/ref|><|det|>[[115, 273, 673, 290]]<|/det|>
+Caption of figure 6 should also more clearly describe what the figures shows.  
+
+<|ref|>text<|/ref|><|det|>[[115, 296, 870, 329]]<|/det|>
+We thank the reviewer for this comment. The captions for revised Fig. 6 and 7 provide additional details about the data presented as compared with the captions in the original manuscript.  
+
+<|ref|>text<|/ref|><|det|>[[115, 344, 866, 377]]<|/det|>
+Minor comments pertain to several typos present in the manuscript like Eukaryotic that should be lower case; Figures are called at time with capital letter and at times with lower case.  
+
+<|ref|>text<|/ref|><|det|>[[116, 384, 285, 400]]<|/det|>
+We corrected the typos.  
+
+<|ref|>text<|/ref|><|det|>[[115, 415, 772, 432]]<|/det|>
+In conclusion I suggest the work to be published once these questions have been addressed.
+
+<--- Page Split --->
+<|ref|>text<|/ref|><|det|>[[119, 84, 293, 97]]<|/det|>
+Reviewers' Comments:  
+
+<|ref|>text<|/ref|><|det|>[[119, 112, 228, 125]]<|/det|>
+Reviewer #1:  
+
+<|ref|>text<|/ref|><|det|>[[119, 127, 298, 139]]<|/det|>
+Remarks to the Author:  
+
+<|ref|>text<|/ref|><|det|>[[119, 140, 835, 168]]<|/det|>
+The authors addressed all major and minor comments listed in the first round of reviews. The manuscript is greatly improved – I would recommend publication.  
+
+<|ref|>text<|/ref|><|det|>[[119, 210, 222, 222]]<|/det|>
+Reviewer #2:  
+
+<|ref|>text<|/ref|><|det|>[[119, 225, 298, 237]]<|/det|>
+Remarks to the Author:  
+
+<|ref|>text<|/ref|><|det|>[[118, 238, 877, 350]]<|/det|>
+First of all, I would like to thank the authors for addressing my comments and questions. I am generally satisfied with their revision but I would need to ask clarification on two points. Supplementary figure 6. The authors state that they are confirming that the expression of BFP does not impact mCherry expression. However, it is not very clear from the figure caption and response in the rebuttal, nor from the methods section, if they actually performed the experiments in presence and absence of competition or if they simply compared the mCherry signal alone with mCherry normalised on BFP but from the same experiment peformed in presence of competition. Can the authors clarify?  
+
+<|ref|>text<|/ref|><|det|>[[118, 350, 860, 378]]<|/det|>
+I would like also to ask the authors if they can comment on the large error bars present for HeLa cell expression experiments.
+
+<--- Page Split --->
+<|ref|>sub_title<|/ref|><|det|>[[115, 90, 200, 102]]<|/det|>
+## Reviewer #2  
+
+<|ref|>text<|/ref|><|det|>[[115, 103, 710, 131]]<|/det|>
+First of all, I would like to thank the authors for addressing my comments and questions. I am generally satisfied with their revision but I would need to ask clarification on two points.  
+
+<|ref|>text<|/ref|><|det|>[[115, 144, 882, 214]]<|/det|>
+Reviewer comment: Supplementary figure 6. The authors state that they are confirming that the expression of BFP does not impact mCherry expression. However, it is not very clear from the figure caption and response in the rebuttal, nor from the methods section, if they actually performed the experiments in presence and absence of competition or if they simply compared the mCherry signal alone with mCherry normalised on BFP but from the same experiment performed in presence of competition. Can the authors clarify?  
+
+<|ref|>text<|/ref|><|det|>[[115, 214, 882, 336]]<|/det|>
+Author response: We thank the reviewer for the opportunity to clarify. We did not create a separate set of clones lacking the BFP gene. The plot presented in Supplementary Fig. 5e (note the renumbering of the supplementary figures) depicts the mCherry channel compared with the mCherry normalized by the BFP channel as measured on the same cells. The purpose of this plot is to show that normalizing by a synthetic house- keeping gene did not affect the trends observed in the data. We remind the reviewer that the utilization of this house- keeping gene was for us to ensure that sorted cells were transfected by the sURS plasmid. In our opinion an experiment lacking this form of "house- keeping" may not be as reliable, since "false- positive" cells may enter the analysis particularly for weakly expressing strains. We have added a statement regarding this experimental strategy to the Figure caption of supplementary figure 6 to alleviate any further confusion.  
+
+<|ref|>text<|/ref|><|det|>[[115, 348, 881, 377]]<|/det|>
+Reviewer comment: I would like also to ask the authors if they can comment on the large error bars present for HeLa cell expression experiments.  
+
+<|ref|>text<|/ref|><|det|>[[115, 376, 882, 459]]<|/det|>
+Author response: The error bar observed for the HeLa cells (i.e. Supplementary Fig. 5h) reflect the actual natural deviation in the expression data observed over the triplicates. The unusual variation may be due to the fact that the intensity of expression was not as strong as compared with the expression measured for the CHO cells, which may have led to increased noise. Fortunately, this expression level was sufficiently strong to differentiate between the different boosts enabled by the sURS variants, which correlated well with both the CHO cell measurements and the revised model.
+
+<--- Page Split --->

peer_reviews/12c8cd2ab4e7cfd26a2e2ccf93098324824525e6e4e9cfa53534dc8d649ac45f/supplementary_0_Transparent Peer Review file/images_list.json

@@ -0,0 +1,430 @@
+[
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_0.jpg",
+    "caption": "Fig. S28. (a) The average hydrodynamic diameter (Z-average) of TBmA aggregates measured by Dynamic Light Scattering (DLS). The distribution of TBmA aggregates during 30 min light irradiation and 72 h FBS preservation.",
+    "footnote": [],
+    "bbox": [
+      [
+        152,
+        108,
+        850,
+        400
+      ]
+    ],
+    "page_idx": 13
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_1.jpg",
+    "caption": "Fig. S31. (a) The average hydrodynamic diameter (Z-average) of TBmA aggregates produced in GGT catalytic reaction measured by DLS. (b-g) Distribution of TBmA aggregates formed at different times of GGT catalytic reaction. (h) The transmission electron microscope (TEM) of the TBmA aggregates formed after the GGT catalytic reaction for 12 h.",
+    "footnote": [],
+    "bbox": [
+      [
+        157,
+        495,
+        848,
+        768
+      ]
+    ],
+    "page_idx": 13
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_2.jpg",
+    "caption": "Fig. S27. The emission wavelength analysis of the LED light.",
+    "footnote": [],
+    "bbox": [
+      [
+        350,
+        664,
+        644,
+        878
+      ]
+    ],
+    "page_idx": 15
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_3.jpg",
+    "caption": "Changes in the Supporting Information: Fig. S35. The anticancer activity ( \\(\\mathrm{IC}_{50}\\) , \\(\\mu \\mathrm{M}\\) ) of TBmA-Glu against GGT overexpressing OVCAR5 and 4T1, and GGT normally expressing HLF1 cells.",
+    "footnote": [],
+    "bbox": [
+      [
+        150,
+        357,
+        866,
+        696
+      ]
+    ],
+    "page_idx": 19
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_4.jpg",
+    "caption": "Fig. S28. (a) The average hydrodynamic diameter (Z-average) of TBmA aggregates measured by Dynamic Light Scattering (DLS). The distribution of TBmA aggregates during 30 min light irradiation (b) and 72h FBS preservation (c).",
+    "footnote": [],
+    "bbox": [
+      [
+        150,
+        492,
+        864,
+        793
+      ]
+    ],
+    "page_idx": 20
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_5.jpg",
+    "caption": "Fig. S34. The HepG2 cells were imaged after incubation with TBmA-Glu (5 \\(\\mu \\mathrm{M}\\) ) for 12 hours. Subsequently, the cells were exposed to a \\(465\\mathrm{nm}\\) laser for 10 minutes, and images were captured every minute. \\(\\lambda_{\\mathrm{ex}} = 465\\mathrm{nm}\\) ; \\(\\lambda_{\\mathrm{em}} = 700\\pm 20\\mathrm{nm}\\) , Scale bar, \\(20\\mu \\mathrm{m}\\) .",
+    "footnote": [],
+    "bbox": [
+      [
+        157,
+        88,
+        864,
+        244
+      ]
+    ],
+    "page_idx": 21
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_6.jpg",
+    "caption": "Changes in the Supporting Information:",
+    "footnote": [],
+    "bbox": [
+      [
+        152,
+        359,
+        866,
+        639
+      ]
+    ],
+    "page_idx": 22
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_7.jpg",
+    "caption": "Fig. S32. Fluorescence emission changes of DCFH (10 \\(\\mu \\mathrm{M}\\) ) in the presence of \\(5\\mu \\mathrm{M}\\) photosensitizers in DMSO-PBS \\((v:v = 1 / 99)\\) after irradiation \\((20\\mathrm{mW}\\cdot \\mathrm{cm}^{-2})\\) for different time. (a) TBmA aggregates in PBS, (b) TBmA aggregates produced in GGT catalytic reaction, DCHF, \\(\\lambda_{\\mathrm{ex}} = 488 \\mathrm{nm}\\) . (c) Plot of the relative emission intensity \\((II_{0})\\) of DCF (10 \\(\\mu \\mathrm{M}\\) ) in presence of TBmA (5 \\(\\mu \\mathrm{M}\\) ), TBmA aggregates produced in GGT reaction (5 \\(\\mu \\mathrm{M}\\) ) or Rose Bengal (RB, \\(5\\mu \\mathrm{M}\\) ) versus the irradiation \\((20\\mathrm{mW}\\cdot \\mathrm{cm}^{-2})\\) time, where \\(I_{0} = \\mathrm{PL}\\) intensity of DCFH in solutions with different water fraction \\((f_{\\mathrm{w}})\\) without light irradiation. \\(\\lambda_{\\mathrm{ex}} = 488 \\mathrm{nm}\\) .",
+    "footnote": [],
+    "bbox": [
+      [
+        157,
+        88,
+        866,
+        249
+      ]
+    ],
+    "page_idx": 23
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_8.jpg",
+    "caption": "Fig. S33. The time-dependent uptake process of TBmA-Glu (a) / TBmA (b) in HepG2 cells. (c) The time-dependent uptake process of TBmA-Glu in LO2 cells. All the cells were incubated with \\(5\\mu \\mathrm{M}\\) TBmA-Glu / TBmA and imaged at the indicated time. \\(\\lambda_{\\mathrm{ex}} = 465 \\mathrm{nm}\\) ; \\(\\lambda_{\\mathrm{em}} = 700 \\pm 20 \\mathrm{nm}\\) , Scale bar, \\(20\\mu \\mathrm{m}\\) .",
+    "footnote": [],
+    "bbox": [
+      [
+        160,
+        87,
+        866,
+        585
+      ]
+    ],
+    "page_idx": 24
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_9.jpg",
+    "caption": "Fig. S28. (a) The average hydrodynamic diameter (Z-average) of TBmA aggregates measured by Dynamic Light Scattering (DLS). The distribution of TBmA aggregates during 30 min light irradiation and 72h FBS preservation.",
+    "footnote": [],
+    "bbox": [],
+    "page_idx": 26
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_10.jpg",
+    "caption": "Fig. S34. The HepG2 cells were imaged after incubation with TBmA-Glu (5 \\(\\mu \\mathrm{M}\\) ) for 12 hours. Subsequently, the cells were exposed to a \\(465 \\mathrm{nm}\\) laser for 10 minutes, and images were captured every minute. \\(\\lambda_{\\mathrm{ex}} = 465 \\mathrm{nm}\\) ; \\(\\lambda_{\\mathrm{em}} = 700 \\pm 20 \\mathrm{nm}\\) , Scale bar, \\(20 \\mu \\mathrm{m}\\) .",
+    "footnote": [],
+    "bbox": [
+      [
+        155,
+        490,
+        864,
+        648
+      ]
+    ],
+    "page_idx": 27
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_11.jpg",
+    "caption": "Fig. S39. Malondialdehyde (MDA) levels in HepG2 cells after treated with the TBmA-Glu (2 \\(\\mu \\mathrm{M}\\) ) for \\(12\\mathrm{h}\\) . Then, the cells were irradiated with a white laser array \\((12\\mathrm{J}\\cdot \\mathrm{cm}^{-2})\\) and the MDA levels were detected using a Lipid Peroxidation MDA Assay Kit. Data expressed as average \\(\\pm\\) standard error, \\(\\mathrm{n} = 3\\) . Statistical significance: P values, \\(\\mathrm{***P}< 0.001\\) , calculated with the Student's T-test.",
+    "footnote": [],
+    "bbox": [
+      [
+        355,
+        421,
+        641,
+        612
+      ]
+    ],
+    "page_idx": 27
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_12.jpg",
+    "caption": "Fig. S36. The effects of hypoxia (2% \\(O_2\\) ) and normoxia (20% \\(O_2\\) ) conditions on the anticancer photodynamic efficiency of Rose Bengal against HepG2 cells.",
+    "footnote": [],
+    "bbox": [
+      [
+        333,
+        83,
+        660,
+        303
+      ]
+    ],
+    "page_idx": 28
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_13.jpg",
+    "caption": "Fig. R1. The long-term stability of TBmA aggregates in \\(30\\%\\) BSA solutions.",
+    "footnote": [],
+    "bbox": [
+      [
+        286,
+        388,
+        707,
+        707
+      ]
+    ],
+    "page_idx": 31
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_14.jpg",
+    "caption": "Fig. R2. The ROS generation capacity of \\(\\mathrm{TBmA}\\) aggregates after dispersed in \\(30\\%\\) BSA solution for \\(0\\mathrm{h}\\) (a) and \\(72\\mathrm{h}\\) (b). The ROS was identified using DCFH as an indicator. (c) The plot of the relative emission intensity \\((I / I_0)\\) of DC versus the irradiation \\((20\\mathrm{mW}\\cdot \\mathrm{cm}^{-2})\\) time, where \\(I_0 = \\mathrm{PL}\\) intensity of DCFH in solutions without light irradiation.",
+    "footnote": [],
+    "bbox": [
+      [
+        160,
+        87,
+        845,
+        250
+      ]
+    ],
+    "page_idx": 32
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_15.jpg",
+    "caption": "Fig. R3 The impact of white light and \\(450\\mathrm{nm}\\) light exposure \\((12\\mathrm{J / cm^2})\\) on the cellular viability of HepG2 cells.",
+    "footnote": [],
+    "bbox": [
+      [
+        365,
+        141,
+        627,
+        302
+      ]
+    ],
+    "page_idx": 33
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_16.jpg",
+    "caption": "Fig. R4 Scheme of the photochemical reactions for type I and type II PDT. \\(^{9}\\)",
+    "footnote": [],
+    "bbox": [
+      [
+        147,
+        195,
+        788,
+        375
+      ]
+    ],
+    "page_idx": 35
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_17.jpg",
+    "caption": "Fig. R5 (a) Cellular viability of HepG2 cells in normoxia and anoxia conditions. (b) Fluorescence emission changes of DCFH (Dichlorodihydrofluorescein, \\(10 \\mu \\mathrm{M}\\) ) in the presence of \\(5 \\mu \\mathrm{M}\\) photosensitizers in DMSO-PBS \\((v:v = 1:99)\\) after irradiation (20 \\(\\mathrm{mW}\\cdot \\mathrm{cm}^{-2}\\) ) for a different time under anoxia conditions. (b) \\(\\mathrm{TBmA}\\) , (c) Rose Bengal (RB). DCHF, \\(\\lambda_{\\mathrm{ex}} = 488 \\mathrm{nm}\\) .",
+    "footnote": [],
+    "bbox": [
+      [
+        155,
+        426,
+        863,
+        595
+      ]
+    ],
+    "page_idx": 37
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_18.jpg",
+    "caption": "Fig. R6 Schematic illustration of the \\(^{1}\\mathrm{O}_{2}\\) generation mechanisms by conventional PDT agents (left) and GQDs (right).",
+    "footnote": [],
+    "bbox": [
+      [
+        198,
+        621,
+        789,
+        821
+      ]
+    ],
+    "page_idx": 37
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_19.jpg",
+    "caption": "Changes in the Supporting Information:",
+    "footnote": [],
+    "bbox": [
+      [
+        150,
+        599,
+        850,
+        891
+      ]
+    ],
+    "page_idx": 38
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_20.jpg",
+    "caption": "Fig. S31. (a) The average hydrodynamic diameter (Z-average) of TBmA aggregates produced in GGT catalytic reaction measured by DLS. (b-g) Distribution of TBmA aggregates formed at different times of GGT catalytic reaction. (h) The transmission electron microscope (TEM) of the TBmA aggregates formed after the GGT catalytic reaction for \\(12\\mathrm{h}\\) .",
+    "footnote": [],
+    "bbox": [
+      [
+        156,
+        153,
+        850,
+        428
+      ]
+    ],
+    "page_idx": 44
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_21.jpg",
+    "caption": "Fig. S27. The emission wavelength analysis of the LED light.",
+    "footnote": [],
+    "bbox": [
+      [
+        348,
+        621,
+        647,
+        834
+      ]
+    ],
+    "page_idx": 45
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_22.jpg",
+    "caption": "Fig. S36. The effects of hypoxia (2% O<sub>2</sub>) and normoxia (20% O<sub>2</sub>) conditions on the anticancer photodynamic efficiency of Rose Bengal against HepG2 cells.",
+    "footnote": [],
+    "bbox": [
+      [
+        333,
+        340,
+        660,
+        562
+      ]
+    ],
+    "page_idx": 47
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_23.jpg",
+    "caption": "Fig. R1. The long-term stability of TBmA aggregates in \\(30\\%\\) BSA solutions.",
+    "footnote": [],
+    "bbox": [
+      [
+        286,
+        92,
+        707,
+        410
+      ]
+    ],
+    "page_idx": 52
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_24.jpg",
+    "caption": "Fig. R2. The ROS generation capacity of TBmA aggregates after dispersed in \\(30\\%\\) BSA solution for \\(0 \\mathrm{~h}\\) (a) and \\(72 \\mathrm{~h}\\) (b). The ROS was identified using DCFH as an indicator. (c) The plot of the relative emission intensity \\((I / I_0)\\) of DC versus the irradiation \\((20 \\mathrm{mW} \\cdot \\mathrm{cm}^{-2})\\) time, where \\(I_0 = \\mathrm{PL}\\) intensity of DCFH in solutions without light irradiation.",
+    "footnote": [],
+    "bbox": [
+      [
+        156,
+        456,
+        848,
+        620
+      ]
+    ],
+    "page_idx": 54
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_25.jpg",
+    "caption": "Fig. R3 The impact of white light and \\(450 \\mathrm{nm}\\) light exposure ( \\(12 \\mathrm{J / cm^2}\\) ) on the cellular viability of HepG2 cells.",
+    "footnote": [],
+    "bbox": [
+      [
+        366,
+        554,
+        627,
+        714
+      ]
+    ],
+    "page_idx": 54
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_26.jpg",
+    "caption": "Fig. R4 Scheme of the photochemical reactions for type I and type II PDT. \\(^{9}\\)",
+    "footnote": [],
+    "bbox": [
+      [
+        148,
+        675,
+        789,
+        857
+      ]
+    ],
+    "page_idx": 56
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_27.jpg",
+    "caption": "Fig. R5 (a) Cellular viability of HepG2 cells in normoxia and anoxia conditions. (b) Fluorescence emission changes of DCFH (Dichlorodihydrofluorescein, \\(10~\\mu \\mathrm{M}\\) ) in the presence of \\(5~\\mu \\mathrm{M}\\) photosensitizers in DMSO-PBS \\((v:v = 1:99)\\) after irradiation ( \\(20\\mathrm{mW}\\cdot \\mathrm{cm}^{-2}\\) ) for a different time under anoxia conditions. (b) \\(\\mathrm{TBmA}\\) , (c) Rose Bengal (RB). DCHF, \\(\\lambda_{\\mathrm{ex}} = 488 \\mathrm{nm}\\) .",
+    "footnote": [],
+    "bbox": [
+      [
+        155,
+        93,
+        864,
+        262
+      ]
+    ],
+    "page_idx": 58
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_28.jpg",
+    "caption": "Fig. R6 Schematic illustration of the \\(^{1}\\mathrm{O}_{2}\\) generation mechanisms by conventional PDT agents (left) and GQDs (right).",
+    "footnote": [],
+    "bbox": [
+      [
+        199,
+        345,
+        789,
+        544
+      ]
+    ],
+    "page_idx": 59
+  }
+]

peer_reviews/12d14f46911ac862fa0888c9a21b38d01db99f055a47f952b75adc66b6066d87/supplementary_0_Peer Review File/images_list.json

@@ -0,0 +1,79 @@
+[
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_0.jpg",
+    "caption": "a: Venn diagram shows the shared protein between PrE genes and SALL4 up-regulated proteins after BMP4 treatment; b: Line chart shows Slc9a3r1 expression pattern in BMP4+ and BMP4- condition; c: Line chart shows Slc9a3r1 expression pattern in WT-SALL4 and delN12-SALL4; d, e, f, g: Histogram shows Oct4 GFP positive iPS colonies numbers in different group, data are mean ± s.d., two-sided, unpaired t test; n = 3 independent experiments, \\(*p < 0.05\\) , \\(**p < 0.01\\) , \\(***p < 0.001\\) .",
+    "footnote": [],
+    "bbox": [],
+    "page_idx": 0
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_1.jpg",
+    "caption": "a, Histogram shows Oct4 GFP positive iPS colonies numbers in different group, data are mean \\(\\pm\\) s.d., two-sided, unpaired t test; \\(n = 3\\) independent experiments, \\(^{*}p< 0.05\\) , \\(^{**}p< 0.01\\) , \\(^{**}p< 0.001\\) ; b: Pictures show the in situ whole well screening of Oct4 GFP positive clone number of different group, scale bar \\(= 5mm\\)",
+    "footnote": [],
+    "bbox": [],
+    "page_idx": 5
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_2.jpg",
+    "caption": "a, b : Histograms show the qPCR results of PrE gene relative expression level of every group, \\(n = 3\\) independent experiments.",
+    "footnote": [],
+    "bbox": [
+      [
+        220,
+        95,
+        714,
+        404
+      ]
+    ],
+    "page_idx": 6
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_3.jpg",
+    "caption": "a: UMAP shows the cells distribution of BMP4+ and BMP4- conditions; b: Stacked barplot shows the cell proportion in each cell type.",
+    "footnote": [],
+    "bbox": [
+      [
+        210,
+        630,
+        785,
+        784
+      ]
+    ],
+    "page_idx": 6
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_4.jpg",
+    "caption": "a: Heatmap shows the common and different features between MEF and BMP4+ cells in cluster 2; b: Bar plots show the top 5 GO terms of gene sets; c: UMAP shows the expression of selected marker genes",
+    "footnote": [],
+    "bbox": [
+      [
+        183,
+        512,
+        792,
+        864
+      ]
+    ],
+    "page_idx": 7
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_5.jpg",
+    "caption": "a, b: UMAP shows the integration of PrECLCs/Pluripotent cells' data in this study and Epiblast/PrE data of Sala2019 in vivo; c, d, e, f: UMAP shows the expression of marker genes at single cell level; g: Correlation heatmap of reprogramming cells and developmental cells.",
+    "footnote": [],
+    "bbox": [],
+    "page_idx": 9
+  },
+  {
+    "type": "image",
+    "img_path": "images/Figure_unknown_6.jpg",
+    "caption": "a, b: Histograms show the qPCR results of PrE gene relative expression level of every group, \\(n = 3\\) independent experiments.",
+    "footnote": [],
+    "bbox": [],
+    "page_idx": 10
+  }
+]