Genomic variation, population history and within-archipelago adaptation between island bird populations

Oceanic island archipelagos provide excellent models to understand evolutionary processes. Colonization events and gene flow can interact with selection to shape genetic variation at different spatial scales. Landscape-scale variation in biotic and abiotic factors may drive fine-scale selection within islands, while long-term evolutionary processes may drive divergence between distantly related populations. Here, we examine patterns of population history and selection between recently diverged populations of the Berthelot's pipit (Anthus berthelotii), a passerine endemic to three North Atlantic archipelagos. First, we use demographic trees and f3 statistics to show that genome-wide divergence across the species range is largely shaped by colonization and bottlenecks, with evidence of very weak gene flow between populations. Then, using a genome scan approach, we identify signatures of divergent selection within archipelagos at single nucleotide polymorphisms (SNPs) in genes potentially associated with craniofacial development and DNA repair. We did not detect within-archipelago selection at the same SNPs as were detected previously at broader spatial scales between archipelagos, but did identify signatures of selection at loci associated with similar biological functions. These findings suggest that similar ecological factors may repeatedly drive selection between recently separated populations, as well as at broad spatial scales across varied landscapes.

1. Please provide some more details on the RAD-data, e.g. coverage per individual and software used. I think that these minimal details should be provided here rather than referring to .
2. You mention reduced genetic diversity on the smaller islands several times in the manuscript but never estimated π. This should be possible with your RAD-data and would be a nice addition.
3. EigenGWAS: I never used that program but as far as I understand it takes the PC1 scores of each individual as the dependent variable and each SNP (used for the PCA) as the independent Decision letter (RSOS-201146.R0) We hope you are keeping well at this difficult and unusual time. We continue to value your support of the journal in these challenging circumstances. If Royal Society Open Science can assist you at all, please don't hesitate to let us know at the email address below.

Dear Ms Martin
The Editors assigned to your paper RSOS-201146 "Genomic variation, population history and within-archipelago adaptation between island bird populations" have now received comments from reviewers and would like you to revise the paper in accordance with the reviewer comments and any comments from the Editors. Please note this decision does not guarantee eventual acceptance.
We invite you to respond to the comments supplied below and revise your manuscript. Below the referees' and Editors' comments (where applicable) we provide additional requirements. Final acceptance of your manuscript is dependent on these requirements being met. We provide guidance below to help you prepare your revision.
We do not generally allow multiple rounds of revision so we urge you to make every effort to fully address all of the comments at this stage. If deemed necessary by the Editors, your manuscript will be sent back to one or more of the original reviewers for assessment. If the original reviewers are not available, we may invite new reviewers.
Please submit your revised manuscript and required files (see below) no later than 21 days from today's (ie 29-Sep-2020) date. Note: the ScholarOne system will 'lock' if submission of the revision is attempted 21 or more days after the deadline. If you do not think you will be able to meet this deadline please contact the editorial office immediately.
Please note article processing charges apply to papers accepted for publication in Royal Society Open Science (https://royalsocietypublishing.org/rsos/charges). Charges will also apply to papers transferred to the journal from other Royal Society Publishing journals, as well as papers submitted as part of our collaboration with the Royal Society of Chemistry (https://royalsocietypublishing.org/rsos/chemistry). Fee waivers are available but must be requested when you submit your revision (https://royalsocietypublishing.org/rsos/waivers). I think this is a very interesting paper on genetic variation within archipelago in the Berthelot's pipit, a passerine found in three archipelagos of the North Atlantic (the Canary islands, the Selvagens and the Madeiran archipelago). As far as I can tell, the analyses are well done, except for one major concern that I have regarding the analyses related to linkage disequilibrium (LD) (see major comment). I mostly have small comments and suggestions to make the manuscript a little bit clearer but this is a very polished version already and I'm recommending to accept this paper with minor revisions.
Reviewer: 2 Comments to the Author(s) In your paper "Genomic variation, population history and within-archipelago adaptation between island bird populations" you perform population genetic analyses and search for loci under divergent selection in Berthelot's pipits sampled across three Macaronesia archipelagos. In general, I enjoyed reading your manuscript although it was at times hard to follow given the many analyses performed and the large supplement. I have a few major points that you may want to consider (order is arbitrary and not by importance): 1. Please provide some more details on the RAD-data, e.g. coverage per individual and software used. I think that these minimal details should be provided here rather than referring to .
2. You mention reduced genetic diversity on the smaller islands several times in the manuscript but never estimated π. This should be possible with your RAD-data and would be a nice addition.
3. EigenGWAS: I never used that program but as far as I understand it takes the PC1 scores of each individual as the dependent variable and each SNP (used for the PCA) as the independent variable. I guess that it does not correct for relatedness.? Thus, it should be similar to the method described in Duforet-Frebourg et al. (2016) or used in Knief et al. (2019). It is basically a representation of the loadings, right? Thus, it (A) is not too surprising that the SNPs you detect are also those with the highest FST (Fig. 5). (B) I was wondering if in the Madeiran archipelago the two significant SNPs generate PC1 because they are in high LD (r2 > 0.982). Would it not be more informative to use the LD-pruned SNP set here? Also, I was wondering whether a Bonferroni-corrected significant threshold is conservative here given that SNPs are in LD. The method implemented in Gao et al. (2008) may be useful here. Last, throughout the manuscript you write "SNPs under selection" (except for the Conclusions paragraph). I would rather say "SNPs putatively under selection" because EigenGWAS does not proof selection.
4. Genotype-phenotype associations: Why do you not fit your relatedness matrix as a random effect (and use all individuals, including the related ones)? Accounting for population is not enough to control the type I error rate here. You used lme4 for fitting the LMMs (not described in the Methods section though and only visible in the provided code) but I think that with the Rpackage lme4qtl you should be able to do this (Ziyatdinov et al. 2018; although I have never used that specific package, I know that lme4 can be modified to handle any relatedness matrix: this function was called relmatmm and can be found via Google). I am wondering why you use principal components and not the raw phenotypes in your genotype-phenotype associations. The PCs are hard to interpret (except for PC1, which mostly reflects size) and I would thus suggest that you use the raw phenotypes as dependent variables and -in case you want this -correct for size by fitting PC1 as an independent variable in your model. Also, using all these PCs may look like data dredging. Table S1: Although the supplement is already extensive, I think that you should split that table into two. Table S1A: The part above the dashed line (i.e. which trimming/filtering steps used for which analyses). Table S1B: Put the different populations as columns and provide the total numbers of individuals and SNPs as rows. Add rows for each filtering step and give the number of individuals/SNPs removed (or kept). Consider adding these two tables to the main text.

5.
6. LD-analysis: Please provide details on the local regression model used for fitting the regression lines in the Methods section. In Fig. 2 I was wondering why there is always a "dip" in the regression line between 25 and 50 kb. Does the local regression model provide estimates such that you can statistically show that "baseline" LD is higher in populations on smaller islands (some kind of intercept)? Why do you not show all populations in Fig. 2 (rather than having an additional Fig. S7)? I think that adding the Selvagens to Fig. 2 is necessary. 7. TreeMix: I am not very familiar with this software but to me the branches in Fig. 2A and even more so in Fig. S2A appear very short. However, you base some conclusions (i.e. that TF was colonized first) partly on the branch lengths. Is this justified? 8. Provide line numbers for the next revision. My minor comments are just "approximate" line numbers.
I hope you find these comments helpful and best wishes.
Minor comments: P3L5: "evolutionary processes" P3L13 and throughout: Could you reformulate "very limited evidence"? Do you have little evidence (i.e. there may or may not be gene flow) or little (if any) gene flow? P3L32: Why not use "depends"? P3L38: "Studies on humans" P3L39: "important for our understanding" P3L41: "insights" P4L13: please reformulate "evolutionary significant adaptations". Significant is a word used in statistics. Are there also "evolutionary non-significant adaptations? Likewise "evolutionary important adaptations" (L21) sounds strange.

===PREPARING YOUR MANUSCRIPT===
Your revised paper should include the changes requested by the referees and Editors of your manuscript. You should provide two versions of this manuscript and both versions must be provided in an editable format: one version identifying all the changes that have been made (for instance, in coloured highlight, in bold text, or tracked changes); a 'clean' version of the new manuscript that incorporates the changes made, but does not highlight them. This version will be used for typesetting if your manuscript is accepted. Please ensure that any equations included in the paper are editable text and not embedded images.
Please ensure that you include an acknowledgements' section before your reference list/bibliography. This should acknowledge anyone who assisted with your work, but does not qualify as an author per the guidelines at https://royalsociety.org/journals/ethicspolicies/openness/.
While not essential, it will speed up the preparation of your manuscript proof if accepted if you format your references/bibliography in Vancouver style (please see https://royalsociety.org/journals/authors/author-guidelines/#formatting). You should include DOIs for as many of the references as possible.
If you have been asked to revise the written English in your submission as a condition of publication, you must do so, and you are expected to provide evidence that you have received language editing support. The journal would prefer that you use a professional language editing service and provide a certificate of editing, but a signed letter from a colleague who is a native speaker of English is acceptable. Note the journal has arranged a number of discounts for authors using professional language editing services (https://royalsociety.org/journals/authors/benefits/language-editing/).

===PREPARING YOUR REVISION IN SCHOLARONE===
To revise your manuscript, log into https://mc.manuscriptcentral.com/rsos and enter your Author Centre -this may be accessed by clicking on "Author" in the dark toolbar at the top of the page (just below the journal name). You will find your manuscript listed under "Manuscripts with Decisions". Under "Actions", click on "Create a Revision".
Attach your point-by-point response to referees and Editors at Step 1 'View and respond to decision letter'. This document should be uploaded in an editable file type (.doc or .docx are preferred). This is essential.
Please ensure that you include a summary of your paper at Step 2 'Type, Title, & Abstract'. This should be no more than 100 words to explain to a non-scientific audience the key findings of your research. This will be included in a weekly highlights email circulated by the Royal Society press office to national UK, international, and scientific news outlets to promote your work.

At
Step 3 'File upload' you should include the following files: --Your revised manuscript in editable file format (.doc, .docx, or .tex preferred). You should upload two versions: 1) One version identifying all the changes that have been made (for instance, in coloured highlight, in bold text, or tracked changes); 2) A 'clean' version of the new manuscript that incorporates the changes made, but does not highlight them.
--An individual file of each figure (EPS or print-quality PDF preferred [either format should be produced directly from original creation package], or original software format).
--An editable file of each table (.doc, .docx, .xls, .xlsx, or .csv --If you are requesting a discretionary waiver for the article processing charge, the waiver form must be included at this step.
--If you are providing image files for potential cover images, please upload these at this step, and inform the editorial office you have done so. You must hold the copyright to any image provided.
--A copy of your point-by-point response to referees and Editors. This will expedite the preparation of your proof.

At
Step 6 'Details & comments', you should review and respond to the queries on the electronic submission form. In particular, we would ask that you do the following: --Ensure that your data access statement meets the requirements at https://royalsociety.org/journals/authors/author-guidelines/#data. You should ensure that you cite the dataset in your reference list. If you have deposited data etc in the Dryad repository, please include both the 'For publication' link and 'For review' link at this stage.
--If you are requesting an article processing charge waiver, you must select the relevant waiver option (if requesting a discretionary waiver, the form should have been uploaded at Step 3 'File upload' above).
--If you have uploaded ESM files, please ensure you follow the guidance at https://royalsociety.org/journals/authors/author-guidelines/#supplementary-material to include a suitable title and informative caption. An example of appropriate titling and captioning may be found at https://figshare.com/articles/Table_S2_from_Is_there_a_trade-off_between_peak_performance_and_performance_breadth_across_temperatures_for_aerobic_sc ope_in_teleost_fishes_/3843624.

At
Step 7 'Review & submit', you must view the PDF proof of the manuscript before you will be able to submit the revision. Note: if any parts of the electronic submission form have not been completed, these will be noted by red message boxes.

RSOS-201146.R1 (Revision)
Review form: Reviewer 1 Is the manuscript scientifically sound in its present form? Yes

Are the interpretations and conclusions justified by the results? Yes
Is the language acceptable? Yes

Recommendation?
Accept with minor revision (please list in comments)

Comments to the Author(s)
Please see attached file (Appendix C).

Review form: Reviewer 2
Is the manuscript scientifically sound in its present form? Yes

Do you have any ethical concerns with this paper? No
Have you any concerns about statistical analyses in this paper? No

Recommendation?
Accept with minor revision (please list in comments) Comments to the Author(s) I acted as reviewer # 2 for the initial submission of your manuscript "Genomic variation, population history and within-archipelago adaptation between island bird populations". I think you now improved the manuscript considerably, taking also my comments and suggestions into account. I have only one major comment remaining and some minor edits: 1. LD-based inferences on Ne. I do not think that loess() is the most appropriate way of fitting a LD-decay function to your data. In the end, you do not get any parameters or statistics on which you could base your conclusions. Also, (1) how did you define the span-parameter (and based on what?) and (2) loess() is know to overreact which may explain the dip around 25-50 kb distance. Thus, I would suggest using an appropriate model here, as for example done in Knief et al. (2017): You could fit the function described in Hill & Weir (1988), which is LD.data ~ ((10 + rho * distance) / ((2 + rho * distance) * (11 + rho * distance))) * (1 + ((3 + rho * distance) * (12 + 12 * rho * distance + ( rho * distance)^2)) / (n * (2 + rho * distance) * (11 + rho * distance))) where n is the sample size and roh represents the population recombination parameter (rho =4 * Ne * r) that is going to be estimated.
I hope you find these comments helpful and best wishes.
Minor comments (line numbers refer to the document with track changes): L84: "selection can also be" L87: "may reveal patterns of strong" L95: "adaptations" L108: "including diseases and climatic factors" L132: Do you mean divergent or disruptive selection instead of "differential"? L141: "across the species range" L151: "variation to variation in phenotypic" L173: "and birds were molecularly" L223: "population structure" L228: Here and throughout: Check the tense you are writing in. I think it must be simple past here, i.e. "We implemented" L253: "analyses" L265: Here and throughout: Do you mean "tree topology" instead of "tree topography"? L265: "pipit populations, we computed" L281: "larger, more outbred populations show a rapid" L288: "Ne, and LD at" L290: "calculated" or "estimated" L294: "the R function". How did you set the span parameter? L305: "LD, which enabled us" L306: "implemented in" L309: "individually for an association with these axes" L313: I guess you either mean above the -log10(P) threshold or below the P-value threshold. L318-320: Delete that sentence, it is repetitive with the above. L325: "After having" L329: Please add version numbers L329: "implemented in R" L332: "SNPs, we fitted LMMs […] as dependent variables." L332: Why do you not test all other phenotypic traits, e.g. wing length? L347: "tree, we trimmed" L366: "showed" L375: "explained" and "variance explained" L378: "placed" L388: "especially in the central" L389: "highest" instead of "strongest" L401: "suggested few" L402: "occurred" instead of "exist"? L407: "smallest and most isolated" L411: You could add the half-decay distances here. L455: "6,252 bases downstream" L492: I do not think that "Analysis of the first" is correct here. It is just the plot of the 1st PC vs 2nd PC. L537: "genotype associations" L538: If you do not use all phenotypes then delete them from the methods section. However, I strongly encourage you to present all results, including those for wing length and weight. L546: "putatively under selection" L577: I think it was "head length", not "beak length".? L578: "and for a SNP in a candidate gene" L587: This needs a quantitative assessment, as described in my major point above. L598: "larger and more outbred" L605: What are these previous studies? L640: "putatively under divergent" L654: "may be under selection" L662-663: I think this is not correct. Or did you control for size somehow? L668: "history, there" L682: "under selection in recently diverged" L734: Please provide also the doi/accession number for the raw sequencing data

Decision letter (RSOS-201146.R1)
We hope you are keeping well at this difficult and unusual time. We continue to value your support of the journal in these challenging circumstances. If Royal Society Open Science can assist you at all, please don't hesitate to let us know at the email address below.

Dear Miss Martin
On behalf of the Editors, we are pleased to inform you that your Manuscript RSOS-201146.R1 "Genomic variation, population history and within-archipelago adaptation between island bird populations" has been accepted for publication in Royal Society Open Science subject to minor revision in accordance with the referees' reports. Please find the referees' comments along with any feedback from the Editors below my signature.
We invite you to respond to the comments and revise your manuscript. Below the referees' and Editors' comments (where applicable) we provide additional requirements. Final acceptance of your manuscript is dependent on these requirements being met. We provide guidance below to help you prepare your revision.
Please submit your revised manuscript and required files (see below) no later than 7 days from today's (ie 14-Dec-2020) date. Note: the ScholarOne system will 'lock' if submission of the revision is attempted 7 or more days after the deadline. If you do not think you will be able to meet this deadline please contact the editorial office immediately.
Please note article processing charges apply to papers accepted for publication in Royal Society Open Science (https://royalsocietypublishing.org/rsos/charges). Charges will also apply to papers transferred to the journal from other Royal Society Publishing journals, as well as papers submitted as part of our collaboration with the Royal Society of Chemistry (https://royalsocietypublishing.org/rsos/chemistry). Fee waivers are available but must be requested when you submit your revision (https://royalsocietypublishing.org/rsos/waivers).
Thank you for submitting your manuscript to Royal Society Open Science and we look forward to receiving your revision. If you have any questions at all, please do not hesitate to get in touch.
Kind regards, Royal Society Open Science Editorial Office Royal Society Open Science openscience@royalsociety.org on behalf of Prof Kevin Padian (Subject Editor) openscience@royalsociety.org Associate Editor Comments to Author: The reviewers are much happier with this iteration of your work and we are grateful that you've carefully engaged with their concerns. That said, there are are a number of matters that need addressing before your paper may be considered ready for acceptance. Firstly, there are a couple of relatively major issues you'll need to tackle, along with a range of smaller tweaks (a certain amount of copy-editing, for instance). Secondly, it has been observed that raw sequencing reads do not appear to be available through your Dryad deposition -please ensure that you check this and either make clearer how they are accessible in Dryad, or make them available through an additional alternative repository. Good luck! Reviewer comments to Author: Reviewer: 2 Comments to the Author(s) I acted as reviewer # 2 for the initial submission of your manuscript "Genomic variation, population history and within-archipelago adaptation between island bird populations". I think you now improved the manuscript considerably, taking also my comments and suggestions into account. I have only one major comment remaining and some minor edits: 1. LD-based inferences on Ne. I do not think that loess() is the most appropriate way of fitting a LD-decay function to your data. In the end, you do not get any parameters or statistics on which you could base your conclusions. Also, (1) how did you define the span-parameter (and based on what?) and (2) loess() is know to overreact which may explain the dip around 25-50 kb distance. Thus, I would suggest using an appropriate model here, as for example done in Knief et al. (2017): You could fit the function described in Hill & Weir (1988), which is LD.data ~ ((10 + rho * distance) / ((2 + rho * distance) * (11 + rho * distance))) * (1 + ((3 + rho * distance) * (12 + 12 * rho * distance + ( rho * distance)^2)) / (n * (2 + rho * distance) * (11 + rho * distance))) where n is the sample size and roh represents the population recombination parameter (rho =4 * Ne * r) that is going to be estimated.
I hope you find these comments helpful and best wishes.
Minor comments (line numbers refer to the document with track changes): L84: "selection can also be" L87: "may reveal patterns of strong" L95: "adaptations" L108: "including diseases and climatic factors" L132: Do you mean divergent or disruptive selection instead of "differential"? L141: "across the species range" L151: "variation to variation in phenotypic" L173: "and birds were molecularly" L223: "population structure" L228: Here and throughout: Check the tense you are writing in. I think it must be simple past here, i.e. "We implemented" L253: "analyses" L265: Here and throughout: Do you mean "tree topology" instead of "tree topography"? L265: "pipit populations, we computed" L281: "larger, more outbred populations show a rapid" L288: "Ne, and LD at" L290: "calculated" or "estimated" L294: "the R function". How did you set the span parameter? L305: "LD, which enabled us" L306: "implemented in" L309: "individually for an association with these axes" L313: I guess you either mean above the -log10(P) threshold or below the P-value threshold. L318-320: Delete that sentence, it is repetitive with the above. L325: "After having" L329: Please add version numbers L329: "implemented in R" L332: "SNPs, we fitted LMMs […] as dependent variables." L332: Why do you not test all other phenotypic traits, e.g. wing length? L347: "tree, we trimmed" L366: "showed" L375: "explained" and "variance explained" L378: "placed" L388: "especially in the central" L389: "highest" instead of "strongest" L401: "suggested few" L402: "occurred" instead of "exist"? L407: "smallest and most isolated" L411: You could add the half-decay distances here. L455: "6,252 bases downstream" L492: I do not think that "Analysis of the first" is correct here. It is just the plot of the 1st PC vs 2nd PC. L537: "genotype associations" L538: If you do not use all phenotypes then delete them from the methods section. However, I strongly encourage you to present all results, including those for wing length and weight. L546: "putatively under selection" L577: I think it was "head length", not "beak length".? L578: "and for a SNP in a candidate gene" L587: This needs a quantitative assessment, as described in my major point above. L598: "larger and more outbred" L605: What are these previous studies? L640: "putatively under divergent" L654: "may be under selection" L662-663: I think this is not correct. Or did you control for size somehow? L668: "history, there" L682: "under selection in recently diverged" L734: Please provide also the doi/accession number for the raw sequencing data Your revised paper should include the changes requested by the referees and Editors of your manuscript. You should provide two versions of this manuscript and both versions must be provided in an editable format: one version identifying all the changes that have been made (for instance, in coloured highlight, in bold text, or tracked changes); a 'clean' version of the new manuscript that incorporates the changes made, but does not highlight them. This version will be used for typesetting. Please ensure that any equations included in the paper are editable text and not embedded images.
Please ensure that you include an acknowledgements' section before your reference list/bibliography. This should acknowledge anyone who assisted with your work, but does not qualify as an author per the guidelines at https://royalsociety.org/journals/ethicspolicies/openness/.
While not essential, it will speed up the preparation of your manuscript proof if you format your references/bibliography in Vancouver style (please see https://royalsociety.org/journals/authors/author-guidelines/#formatting). You should include DOIs for as many of the references as possible.
If you have been asked to revise the written English in your submission as a condition of publication, you must do so, and you are expected to provide evidence that you have received language editing support. The journal would prefer that you use a professional language editing service and provide a certificate of editing, but a signed letter from a colleague who is a native speaker of English is acceptable. Note the journal has arranged a number of discounts for authors using professional language editing services (https://royalsociety.org/journals/authors/benefits/language-editing/).

===PREPARING YOUR REVISION IN SCHOLARONE===
To revise your manuscript, log into https://mc.manuscriptcentral.com/rsos and enter your Author Centre -this may be accessed by clicking on "Author" in the dark toolbar at the top of the page (just below the journal name). You will find your manuscript listed under "Manuscripts with Decisions". Under "Actions", click on "Create a Revision".
Attach your point-by-point response to referees and Editors at Step 1 'View and respond to decision letter'. This document should be uploaded in an editable file type (.doc or .docx are preferred). This is essential.
Please ensure that you include a summary of your paper at Step 2 'Type, Title, & Abstract'. This should be no more than 100 words to explain to a non-scientific audience the key findings of your research. This will be included in a weekly highlights email circulated by the Royal Society press office to national UK, international, and scientific news outlets to promote your work.

At
Step 3 'File upload' you should include the following files: --Your revised manuscript in editable file format (.doc, .docx, or .tex preferred). You should upload two versions: 1) One version identifying all the changes that have been made (for instance, in coloured highlight, in bold text, or tracked changes); 2) A 'clean' version of the new manuscript that incorporates the changes made, but does not highlight them. --If you are requesting a discretionary waiver for the article processing charge, the waiver form must be included at this step.
--If you are providing image files for potential cover images, please upload these at this step, and inform the editorial office you have done so. You must hold the copyright to any image provided.
--A copy of your point-by-point response to referees and Editors. This will expedite the preparation of your proof.

At
Step 6 'Details & comments', you should review and respond to the queries on the electronic submission form. In particular, we would ask that you do the following: --Ensure that your data access statement meets the requirements at https://royalsociety.org/journals/authors/author-guidelines/#data. You should ensure that you cite the dataset in your reference list. If you have deposited data etc in the Dryad repository, please only include the 'For publication' link at this stage. You should remove the 'For review' link.
--If you are requesting an article processing charge waiver, you must select the relevant waiver option (if requesting a discretionary waiver, the form should have been uploaded at Step 3 'File upload' above).
--If you have uploaded ESM files, please ensure you follow the guidance at https://royalsociety.org/journals/authors/author-guidelines/#supplementary-material to include a suitable title and informative caption. An example of appropriate titling and captioning may be found at https://figshare.com/articles/Table_S2_from_Is_there_a_trade-off_between_peak_performance_and_performance_breadth_across_temperatures_for_aerobic_sc ope_in_teleost_fishes_/3843624.

At
Step 7 'Review & submit', you must view the PDF proof of the manuscript before you will be able to submit the revision. Note: if any parts of the electronic submission form have not been completed, these will be noted by red message boxes. The editorial office reopened on 4 January 2021. We are working hard to catch up after the festive break. If you need advice or an extension to a deadline, please do not hesitate to let us know --we will continue to be as flexible as possible to accommodate the changing COVID situation. We wish you a happy New Year, and hope 2021 proves to be a better year for everyone.

Dear Miss Martin,
It is a pleasure to accept your manuscript entitled "Genomic variation, population history and within-archipelago adaptation between island bird populations" in its current form for publication in Royal Society Open Science.
You can expect to receive a proof of your article in the near future. Please contact the editorial office (openscience@royalsociety.org) and the production office (openscience_proofs@royalsociety.org) to let us know if you are likely to be away from e-mail contact --if you are going to be away, please nominate a co-author (if available) to manage the proofing process, and ensure they are copied into your email to the journal.
Due to rapid publication and an extremely tight schedule, if comments are not received, your paper may experience a delay in publication.
Please see the Royal Society Publishing guidance on how you may share your accepted author manuscript at https://royalsociety.org/journals/ethics-policies/media-embargo/. I think this is a very interesting paper on genetic variation within archipelago in the Berthelot's pipit, a passerine found in three archipelagos of the North Atlantic (the Canary islands, the Selvagens and the Madeiran archipelago). As far as I can tell, the analyses are well done, except for one major concern that I have regarding the analyses related to linkage disequilibrium (LD) (see major comment). I mostly have small comments and suggestions to make the manuscript a little bit clearer but this is a very polished version already and I'm recommending to accept this paper with minor revisions.
Major comments: I was not convinced by the use of LD analyses to infer demography. This requires more explanations and references for the reader to understand how to interpret the LD -genetic distance plots in terms of demography, namely bottlenecks and founder events. Here are the sentences that would need more explanations and references: P7.L45: "To further understand population demography, including bottlenecks and population size, LD was calculated for each population in Plink" P17.L34: "examine population level patterns of bottlenecks using linkage disequilibrium (LD) decay." P17.L44: "Past colonisation history and associated bottlenecks are reflected in patterns of population level LD" please add references and explanations to help us understand the rest of the sentence "these indicate larger more outbred populations across the Canary Islands, but increasingly stronger signatures of inbreeding and reduced genetic diversity through bottlenecks and on small isolated islands".

I was not convinced by your interpretations in terms of founder events and bottlenecks, in my opinion, this is not shown by your analyses:
P17.L17: "Patterns of genetic diversity are consistent with signatures of founder events and geographic isolation." P19.L30: "revealing that genetic diversity is largely shaped by colonisation events and related bottlenecks" Here is an additional comment on the LD plots: P9.L34: Could you provide a quantification about the shallowness of the curve? What is the relationship between shallowness and demography (population size)?
I still think it is really interesting to link LD patterns with demography, and I am aware this has been done (see for instance the review by Waples et al. 2016: Waples, R.K., Larson, W.A., and Waples, R.S. (2016). Estimating contemporary effective population size in non-model species using linkage disequilibrium across thousands of loci. Heredity 117, 233-240.) but I was not convinced by the LD plots without any models or curve fitting.
Minor comments: P5.L15: "To address these questions, we first use demographic analyses to examine colonisation history and gene flow across the species range" I am not sure that Treemix and the LD plots qualify as demographic analyses.
P6.L4: How did you identify sex-linked loci? P6.L5: Do you have missing data or did you do any filtering based on missing data? Appendix A P6.L60: supplementary methods -typo last sentence "by was dominated" P7.L7: Why do you think there could be gene flow between islands? This would help to justify the use of Treemix.
P7.L20 "within population": it would help to remind what your populations are, i.e. one population per island except in Tenerife with two populations in the lowlands and in the highlands.
P7.L22 "rejecting migration events when the following event did not significantly increase the likelihood of the model": what number would you consider to be a significant increase of the likelihood? P7.L25: I was not convinced by the choice of the number of SNPs in one block, I find the number 45 to be quite specific. Also, the results with a block of 50 SNPs (and 50 seems to be very close to 45) are quite different for the Canary populations i.e. no east-west trend as discussed in the results and discussion.
P8.L4: What is the general principle of EigenGWAS ? What is the measure calculated and how does this relate to selection? What is the link with principal components mentioned in the results?
P8.L19: Please indicate the program that you used for the morphological PCA.
P8.L33: 9,960 SNPs without any filtration seems to be very low. Did you filter by missing data?
P8.L55: I would have liked to see the results of the PCA first before Treemix analyses as it is an exploratory analysis without any model and assumptions.
P9.L12: "long branch length", I don't think this branch is that long compared to the very long branches of the Selvagens and Madeiran archipelagos. P9.L30: "F3 results suggest weak admixture events subsequent to branch divergence may exist between Madeira and Porto Santo and Fuerteventura/Lanzarote and La Graciosa." Maybe you could add that this is consistent with the geographic distances between the islands. It's reassuring that you did not infer some possibility of admixture between very distant islands.
P9.L41: How about presenting a PCA with all three archipelagos together ? P9.L45: I would rather say "PCA of the Madeiran archipelago separated Deserta Grande from Madeira and Porto Santo along the first axis, and separated Porto Santo from Deserta Grande on the second axis, with a weak gradient from Madeira to Deserta Grande to Porto Santo (Fig. 3B)." P9.L57: What is the link between EigenGWAS and the PCs? Please explain this in the methods as well to make the results more understandable.
P10.L24: How close to the SNP 219s24 are the genes WDH1 and GCH1? Please indicate the chromosome here as well.
P12. Figure2A: I would suggest to add colours or bars with colours for archipelagos grouping the three islands of Madeira and with eastern and western and Tenerife groups of the Canaries so that we can more easily read the figure and relate it with the text.
Figure2B: I don't see the link between LD and Treemix analyses so I would not put them in the same figure. Careful, the "shaded band indicating 95% confidence intervals" does not appear on the plots. P15. Figure 4: The Z chromosome is not present here? If not, could you indicate why? What about the chromosomes 1A and 4A? When reading the scale (-log10(p)), it is not clear what quantity is calculated, and this should be explained in the methods (see previous comment on EigenGWAS). P16.l39: Typos "Because ADAM12 has been shown to play a role" and "is therefore a potential candidate" P16.l42: Typo "A principal component analysis" P16.l43: Please show the PCA variable correlation plots to support the relationships between the PCs and the variables mentioned in the text and show the morphological PCA in the supplementary materials.
P16.l48: Are the results the same for the other SNP?
P17.l22: I may have missed something but it seems to me that the sentences "no overlap between candidate SNPs identified from previous analyses at a broader spatial scale i.e. between archipelagos" and p18.L25 "It is worth pointing out that our outlier SNP in the Canary Islands, 219s24, was in the same broad genomic region on chromosome 5 as a set of SNPs previously identified as being under selection and associated with bill length across the range of Berthelot's pipit " are conflicting.
P17.L37: "Tree topography suggests initial divergence of Berthelot's pipit from the tawny pipit may have been to the central or eastern Canary Islands (Fig. S2A)". I don't think we can say that from figure S2A as the branches of the Canary form a polytomy. I don't think you can say that from figure 2A as well as from what I understood it is not rooted.
P18.L8: Please check all the "within" to decide with the italic style.
P18.L29: "Further research, with a higher density of SNPs, is needed to identify the importance of this genomic region for adaptation in the pipit system." Yes I agree! I think you have a relatively low number of SNPs to perform genomic scans compared to what can be done now from whole genome sequences, so it's important to mention that. You may also elaborate on the very few number of outlier loci that you detected. Do you think you could have missed important outlier loci compared to if you had used whole genome sequences? P44. Figure S11: not quoted in the text Genomic variation, population history and within-archipelago adaptation between island bird populations We thank both reviewers for their extensive and insightful comments, and for the time taken to provide such detailed suggestions on this manuscript. Based on suggestions from Reviewer 1 we have provided additional references and context for our biological interpretations of linkage disequilibrium across the Berthelot's pipit populations, which is a commonly used analysis for understanding genetic diversity. To alleviate the concerns of both reviewers, we rerun our TreeMix models of population divergence across the 13 Berthelot's pipit populations, removing inference of a known rooting population and providing models for additional SNP window sizes. These results show patterns of divergence and admixture within and between archipelagos are consistent with our previous inferences. We have also removed the use of morphological PCA for the genotypephenotype association analyses of our morphological candidate gene ADAM12, and simplified the methods and results in this section. As recommended by Reviewer 2, we have now fitted GLMM models for the ADAM12 SNP genotypes to individual morphology measurements. We find head length to be weakly correlated with genotype at the ADAM12 SNPs, which concurs with our previous conclusions from morphological PC4 (head and tarsus length relative to beak length). Overall our conclusions are robust to these additional analyses, and we believe that these changes, alongside the minor edits detailed below, substantially improve this manuscript. Below we provide a point-bypoint response to each reviewer comment.

Reviewer 1
COMMENT: I think this is a very interesting paper on genetic variation within archipelago in the Berthelot's pipit, a passerine found in three archipelagos of the North Atlantic (the Canary islands, the Selvagens and the Madeiran archipelago). As far as I can tell, the analyses are well done, except for one major concern that I have regarding the analyses related to linkage disequilibrium (LD) (see major comment). I mostly have small comments and suggestions to make the manuscript a little bit clearer but this is a very polished version already and I'm recommending to accept this paper with minor revisions.

RESPONSE:
We thank Reviewer 1 for their comments and are pleased that they found the study interesting and overall well conducted. We have edited the manuscript to address these concerns with particular refinement of the methods and conclusions regarding the population linkage disequilibrium analyses, and provide specific details on how we have addressed these comments below.

Major comments:
COMMENT: I was not convinced by the use of LD analyses to infer demography. This requires more explanations and references for the reader to understand how to interpret the LD -genetic distance plots in terms of demography, namely bottlenecks and founder events. Here are the sentences that would need more explanations and references:

RESPONSE:
We have now provided additional explanation as requested. Patterns of linkage disequilibrium (LD) are commonly used to infer population size and bottleneck history in other species (e.g. humans, wolves and dogs) and we are careful now to cite this work throughout the manuscript to support our methods and conclusions. We also note that patterns of population level LD in the pipit are consistent with signatures found using other methods such as microsatellite Appendix B Genomic variation, population history and within-archipelago adaptation between island bird populations markers and approximate Bayesian computation (Spurgin et al. 2014). We present these results here using RAD-seq markers to provide further insight into population size, bottleneck history and genetic diversity that cannot be revealed by the TreeMix analysis and to add population level detail to previous LD analyses conducted across broader spatial scales in this system .
We now provide additional clarity in the areas suggested by the reviewer: P7.L45: "To further understand population demography, including bottlenecks and population size, LD was calculated for each population in Plink" RESPONSE: Now updated to: L305 -316: "LD summarises both mutational and recombination history, whereby larger more outbred populations have rapid decay of LD between genetic markers compared to small inbred populations (Flint-Garcia et al. 2003). Patterns of LD have been used extensively to detect historic fluctuations in population size (N e ) and founder events in humans (Reich et al. 2001;Reich et al. 2009), selectively bred species such as Chinese Merino sheep, Xinjiang type (Liu et al. 2017) and wild species including European grey wolves, Canis lupus, (Pilot et al. 2014) and village dogs, Canis lupus familiaris (Shannon et al. 2015)……… To further understand population demography, including genetic diversity and population size in the Berthelot's pipit, we calculate LD for each island population using Plink." P17.L34: "examine population level patterns of bottlenecks using linkage disequilibrium (LD) decay."

RESPONSE:
We agree that the bottleneck signature is not conclusive at the population level and have rephrased this sentence: L670-671: "examine population level patterns of genetic diversity using linkage disequilibrium (LD) decay." P17.L44: "Past colonisation history and associated bottlenecks are reflected in patterns of population level LD" please add references and explanations to help us understand the rest of the sentence "these indicate larger more outbred populations across the Canary Islands, but increasingly stronger signatures of inbreeding and reduced genetic diversity through bottlenecks and on small isolated islands". RESPONSE: L681 -690 "…… ; sharp LD decay at proximate SNPs and low long range LD indicates larger more outbred populations across the Canary Islands, while shallow decay patterns and high long range LD indicate bottlenecks and/ or inbreeding and reduced genetic diversity (Pilot et al., 2014;Shannon et al., 2015). Our patterns of LD are consistent with reduced genetic diversity in the Madeira and Selvagens archipelagos and support previously reported patterns based on microsatellite and RAD data (see Spurgin et al. 2014;.
COMMENT: I was not convinced by your interpretations in terms of founder events and bottlenecks, in my opinion, this is not shown by your analyses.

RESPONSE:
High levels of LD are the result of low genetic diversity. We have lower genetic diversity in the later colonised populations, which is particularly clear between the archipelagos and has been Genomic variation, population history and within-archipelago adaptation between island bird populations shown by previous analyses (see Spurgin et al. 2014;. The population level patterns of LD support our previous inferences and we now make clearer reference to both our work and other studies that use these methods: P17.L17: "Patterns of genetic diversity are consistent with signatures of founder events and geographic isolation."

RESPONSE:
We expand on this in the next paragraph as want to avoid external references in the first paragraph of the discussion.
P19.L30: "revealing that genetic diversity is largely shaped by colonisation events and related bottlenecks"

RESPONSE:
We have now removed reference to bottlenecks in the conclusion.

COMMENT:
Here is an additional comment on the LD plots: P9.L34: Could you provide a quantification about the shallowness of the curve? What is the relationship between shallowness and demography (population size)? I still think it is really interesting to link LD patterns with demography, and I am aware this has been done (see for instance the review by Waples et al. 2016: Waples, R.K., Larson, W.A., and Waples, R.S. (2016). Estimating contemporary effective population size in non-model species using linkage disequilibrium across thousands of loci. Heredity 117, 233-240.) but I was not convinced by the LD plots without any models or curve fitting.

RESPONSE:
We did fit a local regression model which is represented by the curve in Figure 2B. Sorry that we did not highlight this clearly enough. This should now be clear in the methods as well as in the Figure 2 legend: L317 -318: "We fitted a local weighted linear regression (loess) curve to the r 2 data using R function 'loess', with 95% confidence intervals calculated." We did however notice that the shaded 95% confidence intervals and individual pairwise SNP r 2 values did not appear on the submitted manuscript version of the plot and this has now been rectified. Together these should aid interpretation of these plots. We decided not to quantify shallowness as we believe that visual inspection of these curves is best given the level of error we have shown. We have added clarification to the relationship between shallowness and population size as above L305 -316, including: L312-314: "The relationship between proximate SNPs reflects historic N e and LD at distant SNPs reflects N e in more recent times . "

Minor comments:
COMMENT: P5.L15: "To address these questions, we first use demographic analyses to examine colonisation history and gene flow across the species range" I am not sure that Treemix and the LD plots qualify as demographic analyses.
Genomic variation, population history and within-archipelago adaptation between island bird populations RESPONSE: Changed to: L153: "….we first use population genetic analyses to examine…." COMMENT: P6.L4: How did you identify sex-linked loci?

RESPONSE:
We used the same method to map Z chromosome as we did for the autosomes. RAD loci were mapped to the Zebra Finch genome (Taeniopygia guttata, version 3.2.4; Warren et al. 2010), with the assistance of the Berthelot's pipit reference genome, as detailed in the methods to determine the genomic location of the SNPs.
COMMENT: P6.L5: Do you have missing data or did you do any filtering based on missing data? RESPONSE: These SNPs were already trimmed for missing data and quality as detailed in . We have added clarification on this in the methods: L196 -203: "The initial ddRAD library was generated using the protocol by DaCosta and Sorenson (2014) which assigns RAD reads to samples based on an 8 bp barcode sequence and retains the read with the highest quality score. Loci that could not be confidently genotyped in more than four samples and those where 10% or more calls were missing or ambiguous were treated as missing data in the "Berthelot's" library. The "All Pipits" dataset containing all Berthelot's pipit and tawny pipit samples was filtered to contain SNPs from RAD loci that were successfully genotyped in 100% of individuals, removing loci that contained SNPs with >2 alleles." COMMENT: P6.L60: supplementary methods -typo last sentence "by was dominated" RESPONSE: Changed to "but was dominated by".
COMMENT: P7.L7: Why do you think there could be gene flow between islands? This would help to justify the use of Treemix.

RESPONSE:
We have previously assessed admixture across the 13 island populations of the pipit and see an east-west population structuring across the Canary Islands, alongside low levels of withinarchipelago divergence (see . We therefore think that it is possible that gene flow between closely located islands may be responsible in part for these patterns. We have added clarification of this point: L240 -248: "Strong population structuring exists between archipelagos of the Berthelot's pipit supporting our previous inference of absence of contemporary gene flow at broad scales in this system . We have reported weak east-west population structuring between populations within the Canary Islands ), but it is unknown whether this is due to contemporary gene flow between closely located islands or weak population divergence since colonisation. We implement TreeMix at these different population scales with the aim of further understanding the evolutionary processes behind the patterns of population structuring we see." Genomic variation, population history and within-archipelago adaptation between island bird populations COMMENT: P12. Figure2A: I would suggest to add colours or bars with colours for archipelagos grouping the three islands of Madeira and with eastern and western and Tenerife groups of the Canaries so that we can more easily read the figure and relate it with the text.

RESPONSE:
Good idea, we have coloured the population labels to match the Figure

RESPONSE:
Our aim was to link the results of population history together (LD -population size and bottlenecks, TreeMix -colonisation, gene flow and drift) and then keep the selection-related plots separate. We think this works, but that it is ultimately an editorial decision and are happy to change these to separate plots if preferred in light of this comment. Thank you for highlighting the missing 95% confidence intervals, this has been rectified in the new submission.
COMMENT: P15. Figure 4: The Z chromosome is not present here? If not, could you indicate why? What about the chromosomes 1A and 4A? When reading the scale (-log10(p)), it is not clear what quantity is calculated, and this should be explained in the methods (see previous comment on EigenGWAS).

RESPONSE:
There are markers mapped to all large chromosomes including the Z chromosome which are represented by the points in Figure 4A and B. We have added additional labels for these large chromosomes in the updated plot. Additional information has been provided in the methods on how the EigenGWAS P value is calculated -see major comment from Reviewer 2.
COMMENT: P16.l39: Typos "Because ADAM12 has been shown to play a role" and "is therefore a potential candidate" COMMENT: P17.l22: I may have missed something but it seems to me that the sentences "no overlap between candidate SNPs identified from previous analyses at a broader spatial scale i.e. between archipelagos" and p18.L25 "It is worth pointing out that our outlier SNP in the Canary Islands, 219s24, was in the same broad genomic region on chromosome 5 as a set of SNPs previously identified as being under selection and associated with bill length across the range of Berthelot's pipit " are conflicting.

RESPONSE:
We have clarified the distinction between these results by adding greater detail to the second quote: L734-736: "It is worth pointing out that while we do not detect selection for the same SNPs within archipelagos as between archipelagos, our outlier SNP in the Canary Islands, 219s24, was in the same broad genomic region on chromosome 5 as a set of ……" COMMENT: P17.L37: "Tree topography suggests initial divergence of Berthelot's pipit from the tawny pipit may have been to the central or eastern Canary Islands (Fig. S2A)". I don't think we can say that from figure S2A as the branches of the Canary form a polytomy. I don't think you can say that from figure 2A as well as from what I understood it is not rooted.

RESPONSE:
This is a fair point, this result is tenuous and we have now removed the assumption of tree rooting for the Berthelot's pipit population tree. We have removed references to specific regions of the Canary Islands for initial divergence/ colonisation and instead just make reference to initial colonisation of the Canary Islands from Figure S2A. Patterns of population divergence are consistent when using an unrooted tree across the Berthelot's pipit populations and so our conclusions regarding divergence across the Berthelot's pipit are independent of this inference.
Genomic variation, population history and within-archipelago adaptation between island bird populations COMMENT: P18.L8: Please check all the "within" to decide with the italic style.

RESPONSE:
We agree that this was confusing and have now removed italics throughout.
COMMENT: P18.L29: "Further research, with a higher density of SNPs, is needed to identify the importance of this genomic region for adaptation in the pipit system." Yes I agree! I think you have a relatively low number of SNPs to perform genomic scans compared to what can be done now from whole genome sequences, so it's important to mention that. You may also elaborate on the very few number of outlier loci that you detected. Do you think you could have missed important outlier loci compared to if you had used whole genome sequences?

RESPONSE:
We agree, and work is currently being undertaken using whole genomes to follow up on these regions and to identify further loci under selection at broad scales and between recently separated populations.
L740 -745: "Using our marker set, we detected only four loci under selection in our analyses. We see low levels of genetic variation across Berthelot's pipit populations as a result of colonisation history and bottlenecks but it is also likely that we have no marker coverage in other regions of the genome that are under selection. Therefore, future studies are needed to uncover greater detail on the loci and hence traits that are under selection in this system." COMMENT: P44. Figure S11: not quoted in the text

RESPONSE:
Reference to this figure is now made in the section discussing the genotype-phenotype association of ADAM12 SNPs.
Genomic variation, population history and within-archipelago adaptation between island bird populations Reviewer 2 COMMENT: In your paper "Genomic variation, population history and within-archipelago adaptation between island bird populations" you perform population genetic analyses and search for loci under divergent selection in Berthelot's pipits sampled across three Macaronesia archipelagos. In general, I enjoyed reading your manuscript although it was at times hard to follow given the many analyses performed and the large supplement. I have a few major points that you may want to consider (order is arbitrary and not by importance):

RESPONSE:
We thank Reviewer 2 for their thoughtful and detailed comments on this manuscript. We are glad that the reviewer found the manuscript interesting and have tried to improve the clarity of analysis results as you have mentioned in your comments so that the conclusions may be better understood by the reader. We respond to each comment in turn below.
COMMENT: 1. Please provide some more details on the RAD-data, e.g. coverage per individual and software used. I think that these minimal details should be provided here rather than referring to .

RESPONSE:
Please see response to Reviewer 1 -we have included specific details on the bioinformatics performed to generate the initial RAD libraries used in this study.
COMMENT: 2. You mention reduced genetic diversity on the smaller islands several times in the manuscript but never estimated π. This should be possible with your RAD-data and would be a nice addition.

RESPONSE:
Nucleotide diversity was calculated per SNP and per locus for each of the archipelagos, and separately for Deserta Grande island, in . In the interest of keeping the manuscript succinct we avoided replicating these analyses here with just a few additions at population level. We now make more explicit reference to these previous results to back up our inferences of genetic diversity from the LD analysis. See comments to Reviewer 1 for details.

RESPONSE:
Yes this is correct. We have added additional detail on the theory of EigenGWAS in the methods so this should be easier to follow (see response to Reviewer 1). We remove closely related individuals for both the PCA and EigenGWAS analysis, so that the PCs are not influenced by uneven sampling of individuals across populations. EigenGWAS implements a genomic inflation correction for P values ( GC ) to correct for population stratification (i.e., drift), to avoid interpreting ancestryinformative markers as loci under selection. Then, using these corrected P values we test for SNPs putatively under selection along that PC axis. The Bonferroni corrected significance threshold is conservative, but we believe this is the best approach for exploratory analyses such as these. It is useful (and standard) to view SNPs in context of each other which would not be possible with an LDpruned set.
We agree that it is not surprising that the EigenGWAS results are highly correlated with F ST , as they are both reliant on allele frequencies. We present the F ST results for two reasons. First, many other studies rely solely on F ST to detect such outliers and so we feel it is important to show that both analyses reveal the same outliers. Secondly, it is an interesting visual representation of the fact that very different SNPs are under divergence within the Madeiran archipelago compared to the Canary Islands which is not presented by the EigenGWAS plots.
Finally, we agree that we cannot be confident that these SNPs are under selection and have changed our wording to "SNPs putatively under selection" throughout.
COMMENT: 4. Genotype-phenotype associations: Why do you not fit your relatedness matrix as a random effect (and use all individuals, including the related ones)? Accounting for population is not enough to control the type I error rate here. You used lme4 for fitting the LMMs (not described in the Methods section though and only visible in the provided code) but I think that with the Rpackage lme4qtl you should be able to do this (Ziyatdinov et al. 2018; although I have never used that specific package, I know that lme4 can be modified to handle any relatedness matrix: this function was called relmatmm and can be found via Google). I am wondering why you use principal components and not the raw phenotypes in your genotypephenotype associations. The PCs are hard to interpret (except for PC1, which mostly reflects size) and I would thus suggest that you use the raw phenotypes as dependent variables and -in case you want this -correct for size by fitting PC1 as an independent variable in your model. Also, using all these PCs may look like data dredging.

RESPONSE:
Given the low relatedness values in our filtered dataset, we believe that fitting the relatedness matrix as a random effect is unnecessary and would result in overfitting given our sample size. Further, we have been very careful throughout to emphasise the tentative nature of our results in relation to morphology.
However, we do agree on the PCA, and we have now simplified these analyses to remove the morphology principal components and instead investigate effect of ADAM12 genotype on three morphological measurements of head length, wing length and bill length. We find significant effect of genotype on head length and report results for these three GLMMs.
COMMENT: 5. Table S1: Although the supplement is already extensive, I think that you should split that table into two. Table S1A: The part above the dashed line (i.e. which trimming/filtering steps used for which analyses). Table S1B: Put the different populations as columns and provide the total numbers of individuals and SNPs as rows. Add rows for each filtering step and give the number of individuals/SNPs removed (or kept). Consider adding these two tables to the main text.
I thank the authors for taking into account my suggestions as well as the ones from reviewer 2. I think that the manuscript has been subsequently improved and I am very happy with the answers you provided. However, in my opinion, one concern remains regarding the Linkage Disequilibrium analyses. I find them a lot clearer now that you have provided context and I have carefully read the references that you added which I think are extremely interesting. I agree with you that the Selvagens have a higher baseline LD, as the Madeiran island populations, that could suggest a reduced genetic diversity and small population size. My question concerns what you mean by "shallow" and "sharp". When looking at Figure 2B, I see that the curves of the Madeiran islands (green) and the Selvagens (orange) are sharper than the curves of the Canary Islands (purple), since the initial slope is more negative in the curves of the Madeiran and Selvagen islands than in the curves of the Canary Islands. Yet, you say that "LD decay was shallow across the three Madeiran islands", so I did not understand well what you mean by sharp and shallow. In addition, still concerning the LD analyses, in all the papers that you cited, simulations were conducted to confirm than the shape of the LD curve was consistent with a bottleneck. I think this would be a very interested venue, but as you already did many analyses and the paper is close to publication, maybe you could only mention that a simulation-based approach would help confirming that these LD plots can be inferred in terms of bottlenecks. Finally, another interesting analysis would be to use your r2 estimations to infer population size and variation in population size over time from Sved Equation (1971) as in Liu et al. 2017or Pilot et al. 2014 if you have enough data for this. I'm particularly interested in knowing if you can decipher a long term low population size because the islands are small from a recent bottleneck due to founder event as in Pilot et al. 2014.
Here are some additional minor changes that I'm suggesting. Please note that the line numbers correspond to the manuscript with track changes: L 51. change is dependent to depends upon You removed 'demography' in the conclusion, so should you remove it also here?: L 271. 289.
L 290. we calculated L 374. the Tawny pipit rooted tree obtained without adding migration events L 378. including just  limited to Figure 1: maybe add the three colours of the different archipelagos, for example as a frame in each of the archipelago name? And in that case don't use the green in the zooming frame as it is already used for the Maldeiran archipelago. Figures S9 and S10: indicate the chromosome Z 742. whose I want to congratulate again the authors for this manuscript that I really much enjoyed reading and I hope that they will find my questions about linkage disequilibrium analyses relevant.