This paper reports sequencing, population history inferences, and selective sweep mapping in ducks using whole genome sequence data of multiple populations.

This is a good paper. It presents a large-scale population genomic dataset of ducks, uses standard methods that seem appropriate to the task, and it is well written.

Despite this, I have a few criticisms and questions:

1. The paper repeatedly states that this is the first time MITF is associated with colour in the duck. This seems not to be entirely true (see Li et al 2012, http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0036592, and Sultana et al 2017, https://www.ncbi.nlm.nih.gov/pubmed/28823136, but maybe the latter was not published when the manuscript was written). This study presents a whole-genome scan, which should provide stronger evidence than candidate gene associations. Comparing to other papers would be interesting. Can that help filter the candidate variants?

2. It would be useful to see the population history results put more into context. In the light of what is known about duck breed history, is it reasonable that meat and egg type ducks split 2100 years ago? In the Discussion, this number is said to be "compatible with previous written records from 500 BC". The reference is to a book with no page numbers given. Would it be possible to be more specific? Given convergence problems with alternative models, how sure are you that the balance between migration and split time is right? I will admit that I am not really the person to evaluate the pairwise sequential Markov coalescent and δaδi results.

3. It is nice to see the high overlap between SNPs detected here and those in dbSNP. How many of the indels were already in databases? Was PCR validation only for SNPs? Given that indel detection is harder than SNP detection, are you convinced that the MITF indels are real?

4. A protocol for PCR validation seems to be missing (L440-442). It is hard to interpret the 100% accuracy in SNP validation when it is not clear how validation was performed or the accuracy evaluated.

5. The paper is well written, but the GigaScience author guidelines prescribe a somewhat different structure. It specifies an abstract divided into Background, Results, and Conclusions. The Data Description section is missing and other sections are have different names.

6. It seems to me that the data and source code availability may not be in line with the journal policies. I am not certain how to interpret the policies, but the editors will know better. Overall, the methods are described in text, but protocols and scripts are not provided. The raw sequence data is published in a repository, but little else, not even the full population genetic statistics or location of sweeps, as far as I can tell.

Minor comments

Line 35: The important numbers are the number of individuals sampled and the coverage per individual. Average coverage per breed seems less interesting.

Lines 97-101: What do the average numbers of variants detected per individual mean? Are they variants that differ from reference genome, heterozygous variants, or something else?

Lines 243-250: Which GO terms were these, and how were they chosen? It seems odd to me to first select a subset of genes based on GO and then perform enrichment analysis on that set. Will this not bias the analysis?

Lines 393-400: Is there a reason for this mix of sequencing coverage?

Lines 381-384: It is not clear where the ducks came from. How were they obtained?

Line 506: What tool was used for Fst? Also VCFtools?

Figure 1b: The circos plot in Figure 1 looks impressive, but is impossible to read. What is it supposed to show?

Throughout methods: Version numbers are missing for some softwares.

Are the methods appropriate to the aims of the study, are they well described, and are necessary controls included? If not, please specify what is required in your comments to the authors.
Yes

Are the conclusions adequately supported by the data shown? If not, please explain in your comments to the authors. Yes

Does the manuscript adhere to the journal’s guidelines on minimum standards of reporting? If not, please specify what is required in your comments to the author
No

Are you able to assess all statistics in the manuscript, including the appropriateness of statistical tests used? (If an additional statistical review is recommended, please specify what aspects require further assessment in your comments to the editors.)
Yes, and I have assessed the statistics in my report.

Quality of written English Please indicate the quality of language in the manuscript:
Acceptable

Declaration of competing interests Please complete a declaration of competing interests, consider the following questions: Have you in the past five years received reimbursements, fees, funding, or salary from an organization that may in any way gain or lose financially from the publication of this manuscript, either now or in the future? Do you hold any stocks or shares in an organization that may in any way gain or lose financially from the publication of this manuscript, either now or in the future? Do you hold or are you currently applying for any patents relating to the content of the manuscript? Have you received reimbursements, fees, funding, or salary from an organization that holds or has applied for patents relating to the content of the manuscript? Do you have any other financial competing interests? Do you have any non-financial competing interests in relation to this manuscript? If you can answer no to all of the above, write ‘I declare that I have no competing interests’ below. If your reply is yes to any, please give details below.
I declare that I have no competing interests

I agree to the open peer review policy of the journal. I understand that my name will be included on my report to the authors and, if the manuscript is accepted for publication, my named report including any attachments I upload will be posted on the website along with the authors' responses. I agree for my report to be made available under an Open Access Creative Commons CC-BY license (http://creativecommons.org/licenses/by/4.0/). I understand that any comments which I do not wish to be included in my named report can be included as confidential comments to the editors, which will not be published.
I agree to the open peer review policy of the journal.

Authors' response to reviews: Reviewer #1: This paper reports sequencing, population history inferences, and selective sweep mapping in ducks using whole genome sequence data of multiple populations.

This is a good paper. It presents a large-scale population genomic dataset of ducks, uses standard methods that seem appropriate to the task, and it is well written.

Despite this, I have a few criticisms and questions:

Comment: 1. The paper repeatedly states that this is the first time MITF is associated with colour in the duck. This seems not to be entirely true (see Li et al 2012, http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0036592, and Sultana et al 2017, https://www.ncbi.nlm.nih.gov/pubmed/28823136, but maybe the latter was not published when the manuscript was written). This study presents a whole-genome scan, which should provide stronger evidence than candidate gene associations. Comparing to other papers would be interesting. Can that help filter the candidate variants?

Reply: Thank you very much for your positive comments and for the two very helpful citations. Li et al (2012) identified that M isoform of MITF as expressed in black feather ducks, rather than white feather ducks or other colorful ducks. Sultana et al (2017) showed several SNPs and INDEL of MITF with different allele frequency in black and white ducks (table 2 - 5), but did not distinguish the correlation of MITF to white or other feather colors. Due linkage effects, it is notoriously difficult to determine which variant is the real causative mutation of white plumage. Thus, we used the strictest variant filter criteria, namely those with fixed genotype differences in white and non-white ducks. We would very much like to implement the reviewer’s suggestion of using the variants identified in these two previous studies, however the variants reported in Li et al (2012) and Sultana et al (2017) do not in fact pass our strict filter criteria.

We have however added these citations to our manuscript and revised the discussion accordingly (please see line 390). Most importantly, in order to distinguish our result from these previous studies, we revised our statement to say that “Our results show that white plumage in the duck is completely associated with selection at the MITF locus” in our current manuscript, please see line 42 and line 246-247.

Comment: 2. It would be useful to see the population history results put more into context. In the light of what is known about duck breed history, is it reasonable that meat and egg type ducks split 2100 years ago? In the Discussion, this number is said to be "compatible with previous written records from 500 BC". The reference is to a book with no page numbers given. Would it be possible to be more specific? Given convergence problems with alternative models, how sure are you that the balance between migration and split time is right? I will admit that I am not really the person to evaluate the pairwise sequential Markov coalescent and δaδi results.

Reply: Many thanks for your comments. As we state in the manuscript, written records note domestic ducks in China as early as 500 BC. Due to the lack of archaeological evidence, we must focus on textual evidence, which indicates duck domestication occurred approximately 2,000 - 2,500 years ago. We have added these historical references regarding duck domestication to our current manuscript, please see lines 63-71, and have added page numbers to the book citations, and below, please see lines 697-700. Meanwhile, we also reran the PSMC and δaδi analyses based on the mutation rate estimate in chicken (1.91 x 10-9 per base per generation, Nam et al. 2010). The chicken is phylogenetically closer to the duck than zebra finch, the source of our previous mutation rate estimate (Jarvis et al. 2014), however the mutation rate estimates in both chicken and duck are qualitatively similar. As a result, our results are similar, and indicate duck domestication occurred 2228 (441) years ago. We revised the PSMC and δaδi results of our current manuscript, please see Fig 2D, Table 1, and lines 204-219, 546-548.

It is true that the recent divergent time and the high level of diversity in both the domestic and wild populations makes it difficult to differentiate recent admixture from incomplete lineage sorting, however our genetic analysis is largely consistent with these written records, and does not indicate domestication much earlier than this time.

Luff R. 2000. Ducks. In Cambridge World History of Food, ed. KF Kiple, KC Ornelas, pp. 517–24. Cambridge, UK: Cambridge University Press Jarvis, E. D., et al. (2014). "Whole-genome analyses resolve early branches in the tree of life of modern birds." Science 346(6215): 1320-1331. Nam, K., et al. (2010). "Molecular evolution of genes in avian genomes." Genome Biol 11(6): R68.

Comment: 3. It is nice to see the high overlap between SNPs detected here and those in dbSNP. How many of the indels were already in databases? Was PCR validation only for SNPs? Given that indel detection is harder than SNP detection, are you convinced that the MITF indels are real?

Reply: Thank you for your comments. Initially, we validated our INDELs in dbINDEL, following a similar protocol to our SNP validation. However, there has been less focus on INDEL annotation in the database, which contains nearly 70 fold fewer INDELs than we detected. As we used extremely strict filter criteria for INDELs as well as SNPs, we suggest that the difference in variation is due to our greater focus on INDEL annotation please lines 497 – 500. For the two MITF INDELs discussed, we used diagnostic PCR combined with Sanger sequencing to validate these sites in the 78 white and non-white ducks, as well as the first three SNPs (SNP817793, SNP817818, and SNP818004). The Sanger sequencing results of the three SNPs and INDEL817958 completely match our NGS analysis, please see figure below and supplemental figure S5 in our current manuscript. For INDEL818495, we were unable to identify a suitable PCR primer. We have added this to our revised manuscript, please see lines 247-253.

Comment: 4. A protocol for PCR validation seems to be missing (L440-442). It is hard to interpret the 100% accuracy in SNP validation when it is not clear how validation was performed or the accuracy evaluated.

Reply: Apologies, and many thanks for pointing this out. The SNP validation was performed by diagnostic PCR combined with Sanger sequencing method. We have added this description to our revised manuscript, please see lines 510-513.

Comment: 5. The paper is well written, but the GigaScience author guidelines prescribe a somewhat different structure. It specifies an abstract divided into Background, Results, and Conclusions. The Data Description section is missing and other sections are have different names.

Reply: Thank you very much for this helpful suggestion. We had separated the abstract section accordingly, please see lines 30-44. We have also added the Data Description section, please see lines 86-109. We also renamed the Results as Analyses, please see line 111, and revised the Availability of Supporting Data and Materials (lines 618-628), and the Declarations section (lines 632, 633, and 641).

Comment:6. It seems to me that the data and source code availability may not be in line with the journal policies. I am not certain how to interpret the policies, but the editors will know better. Overall, the methods are described in text, but protocols and scripts are not provided. The raw sequence data is published in a repository, but little else, not even the full population genetic statistics or location of sweeps, as far as I can tell.

Reply: Apologies for our previous raw data and source code status. The data from the 78 ducks used in whole genome resequencing and the 14 ducks used in RNA-seq analysis have been submitted to NCBI BioProject (http://www.ncbi.nlm.nih.gov/bioproject) under accession numbers PRJNA419832 and PRJNA419583, respectively. The unassessembled sequencing reads of 78 ducks and RNA-seq reads of 14 ducks have been deposited in the NCBI Sequence Read Archive (SRA: http://www.ncbi.nlm.nih.gov/sra) under accession numbers SRP125660 and SRP125529, respectively. VCF files of SNPs and INDELs, as well as other supporting data, have been submitted to GigaDB as you suggest, please check the GigaDB servers. And, we add these description to our current manuscript, please see lines 618-628.

Minor comments

Comment: Line 35: The important numbers are the number of individuals sampled and the coverage per individual. Average coverage per breed seems less interesting.

Reply: Many thanks for your comment, we had revised this to per individual coverage information, please see line 36.

Comment: Lines 97-101: What do the average numbers of variants detected per individual mean? Are they variants that differ from reference genome, heterozygous variants, or something else?

Reply: Many thanks for your questions. The number of variants between the reference genome and each individual are different, especially in wild mallard and domesticated ducks, (please see supplementary table S2). The average value is the mean variant count of an individual, which includes both heterozygous variants and homozygous variants.

Comment: Lines 243-250: Which GO terms were these, and how were they chosen? It seems odd to me to first select a subset of genes based on GO and then perform enrichment analysis on that set. Will this not bias the analysis?

Reply: Apologies for any confusion. In fact, we observed 292 genes in the top 5% Fst regions, please see supplementary table S5. Our enrichment analysis is based on these 292 genes, and we identified a subset of GO terms for further analyses based on significant GO term P-values, please see supplementary table S7. Moreover, we add the full GO terms to our current manuscript, please see supplementary table S6.

Comment: Lines 393-400: Is there a reason for this mix of sequencing coverage?

Reply: We aimed to sequence each individual at 5X coverage. Additionally, in order to reduce the false negative rate of variants due to our strict filter criteria, we randomly selected one individual from each population for 10X coverage.

Comment: Lines 381-384: It is not clear where the ducks came from. How were they obtained?

Reply: Many thanks for your questions. PK and ML ducks were obtained from Institute of Pekin Duck with the help of Mr. Fangxi Yang, please see author information section, lines 5 and 25. CV ducks were obtained from Cherry Valley farms Co. Ltd with the help of Dr. Yong He, please see lines 5 and 26. The other domesticated ducks were obtained from different duck breeding farms under the help of Dr. Huifang Li, please see lines 5 and 23.

Comment: Line 506: What tool was used for Fst? Also VCFtools?

Reply: Thanks you very much for your questions. The Fst was calculated by the formula described by Weir BS (1984) under our custom perl script. Our custom perl script have been submitted to GigaDB database.

Weir, B. S. and C. C. Cockerham (1984). "Estimating F-Statistics for the Analysis of Population-Structure." Evolution 38(6): 1358-1370.

Comment: Figure 1b: The circos plot in Figure 1 looks impressive, but is impossible to read. What is it supposed to show?

Reply: Apologies for any problems with our figures. The complicated circos plot is the result of the many scaffolds (78,488) in the current duck reference genome. We have removed the circos plot from our current manuscript, please see figure 1, and line 125-127 .

Comment: Throughout methods: Version numbers are missing for some softwares.

Reply: Apologies for this. We have added all this information to our current manuscript, such as NGS QC Toolkit v2.3.3 (line 480), SnpEff v4.0 (line 501), GCTA v1.25 (line 520), MUSCLE v3.8 (line 532), PSMC v0.6.5 (line 541), ∂a∂i v1.7 (line 550), VCFtools v0.1.13 (line 592), and edgeR v3.6 (line 617).

Reviewer #2:

Zhang et al. sequenced whole genomes of 78 individuals of domesticated and wild mallard populations. The authors find a complex history of domestication, with particular artificial selection of meat and egg production in domesticated lineages. Further, outlier analyses demonstrate that white plumage was the result of selection of MITF transcriptional factors. I believe that the authors are tackling an important question regarding variation between domesticates and wild populations, and with an extensive genomic dataset. However, I think the authors fall short in introducing the subject and discussing their results. Moreover, the manuscript requires editing prior to publication, particularly the introduction.

Comment: Introduction. The introduction requires extensive editing. I would also encourage the authors to add another sentence as the relevance (the why) of looking for outliers between domesticated and wild stocks. What exactly are you trying to learn? Instead of results, I would like to see hypotheses regarding what the authors may expect when comparing the genomes of domesticated and wild populations.

Reply: Many thanks for your comments. The most important reason we identified outliers between wild and domesticated ducks was to identify putative sites associates with the genetic basis of phenotypic differences between wild and domestic populations. We have added this explanation to our manuscript, and have also extensively revised our introduction section according to your suggestions, please see lines 51-85.

We had two primary hypotheses regarding duck domestication given the deep divergence between meat and egg breeds. Were ducks domesticated once from wild mallards and subsequently selected for separate egg and meat traits, or were egg and meat populations domesticated in two independent events. We have add the hypotheses of duck domestication scenarios to introduction section, please see lines 75-79.

Comment: The whole first paragraph requires editing. For example -- Line 50-52: Suggest change sentence to: "Mallards (Anas platyrhynchos) are the world's most widely distributed and agriculturally important waterfowl species, and are especially of economic importance in Asia [1]."

Reply: Many thanks for this suggestion. We had revised the sentence accordingly, please see lines 63-64. And we have also extensively revised the first paragraph as suggested, please see lines 52-71.

Comment: Results 1. Line 79 - is this 535 billion mappable reads per sample or across samples?

Reply: Apologies for any confusion. The 535 billion is the total mapped reads across samples. We have added this explanation to our revised manuscript, please see line 117.

Comment: 2. Lines 115-121- how did the authors pick the optimum K in FRAPPE analyses? Did the authors explore additional K values? Where separate analyses done within wild and domesticated populations? Please explain.

Reply: Many thanks for your comments. We analyzed the population structure with K =2, 3 and 4 because there are four duck types across the nine duck populations, shown below, and explained in lines161-165. When K=4, a clear division was found between egg type ducks (JD, SM, and SX) and dual-purpose type ducks (GY) (supplemental figure S6). The most important reason we focused on K=3 as the optimum value for further analysis is due to the results of both the phylogenic and PCA analyses, which convergently showed the nine duck populations clustered into 3 major groups.

Comment: 2a. What do the authors make of domesticated admixture in wild populations? Is this hybridization, ancestry, a combination of both…? I would encourage the authors to explore this further as hybridization between domesticated and wild breeds is a serious concern for conservation of wild populations.

Reply: We agree with the reviewer that this is a very interesting area, and an area of great conservation importance. Unfortunately, given the recent domestication and high levels of diversity we observe, it is not in fact possible to accurately differentiate hybridization from incomplete lineage sorting with our current data, as complex models with these alternative scenarios failed to converge. We agree that this is an interesting area for further study, and have added this explanation to our current manuscript, please see lines 377-381.

Comment: 2b. The PCA analyses seem to suggest that there is structure within wild populations. Running a FRAPPE analyses on wild populations could help tease out whether they are 1 population and PCA analyses are just separating samples as there is so much variation.

Reply: Thank you very much for your comments. Of course, the PCA result showed there is a structure within wild populations, because the two wild populations come from two different provinces in China separated by nearly 2,000 km, (please see line 446). However, the PCA result also showed extensive overlap of these two wild populations, please see fig 2B. Additionally, our FRAPPE analyses were based on all 78 duck individuals rather than pooled population information. Thus, we apologize if we have missed something intended by the reviewer, but we think the structural analysis suggested with recover the same result as our current analysis.

Comment: 3. Lines 139-141 - consider revising the sentence into a more formal hypothesis. I would also like to see such hypotheses in the introduction.

Reply: Thank you so much for your kind suggestion. We had two primary hypotheses regarding duck domestication given the deep divergence between meat and egg breeds. Were ducks domesticated once from wild mallards and subsequently selected for separate egg and meat traits, or were egg and meat populations domesticated in two independent events. We have added the hypotheses of duck domestication scenarios to introduction section, please see lines 75-79.

Comment: 4. Outside of outlier tests by calculating FST, the authors should consider more formal testing of these putative outliers (e.g., BayeScan).

Reply: Thank you very much for this suggestion. We have recalculated our FST with BayeScan, and the results are statistically similar to our current analysis, based on Weir, B. S. (1984). Thus, we have kept our previous FST method in our revised manuscript, as this method is a classical and formal method for calculating FST, and has been widely implemented in many organisms, including rice (Meyer, R. S., et al. 2016), sheep (Yang, J., et al. 2016), dog (Gou, X., et al. 2014, Axelsson, E., et al. 2013), and pigeon (Shapiro, M. D., et al. 2013).

Weir, B. S. and C. C. Cockerham (1984). "Estimating F-Statistics for the Analysis of Population-Structure." Evolution 38(6): 1358-1370. Meyer, R. S., et al. (2016). "Domestication history and geographical adaptation inferred from a SNP map of African rice." Nat Genet 48(9): 1083-1088. Yang, J., et al. (2016). "Whole-Genome Sequencing of Native Sheep Provides Insights into Rapid Adaptations to Extreme Environments." Mol Biol Evol 33(10): 2576-2592. Gou, X., et al. (2014). "Whole-genome sequencing of six dog breeds from continuous altitudes reveals adaptation to high-altitude hypoxia." Genome Res 24(8): 1308-1315. Axelsson, E., et al. (2013). "The genomic signature of dog domestication reveals adaptation to a starch-rich diet." Nature 495(7441): 360-364. Shapiro, M. D., et al. (2013). "Genomic diversity and evolution of the head crest in the rock pigeon." Science 339(6123): 1063-1067.

Comment: 5. Although I like the idea of RNA-seq data here. I think that this is largely overlooked in the manuscript and may detract from the main (genome) focus. I would encourage the authors to consider taking the RNA-seq out or sufficiently expanding on methods, reasoning, etc. of the RNA-seq data.

Reply: Thank you so much for your suggestion. We respectfully suggest that the RNA-seq is a key component of our manuscript, as it represents functional phenotypic differentiation of wild mallards and domesticated ducks, and helps connect the genomic variation to phenotypic differences. We have revised the methods and reasoning of including this data RNA-seq as suggested, please see lines 324-328, 470-475, and 603-615.

Comment: 6. I would like to see global Fst estimates among breeds, wild locations

Reply: Many thanks for your comment. The global FST between were showed in below, and we also add this table to our current manuscript, please see lines 267-268, and supplemental table S4.

Comment: Discussion I have no issues with the discussion and find it the best written. I think that a section on domesticate and wild hybridization may broaden the appeal of this paper.

Reply: Thanks for this suggestion. As we mentioned above, given the recent domestication and high levels of diversity we observe, it is not possible to accurately differentiate hybridization from incomplete lineage sorting with our current data, as complex models with these alternative scenarios failed to converge. We agree that this is an interesting area for further study, and have added material to the discussion as suggested, please see lines 377-381.

Comment: Methods Please add additional information regarding FRAPPE analyses, K selection,etc.

Reply: Apologies for any omissions. We have added the method of FRAPPE analyses and K selection to our current manuscript, please see lines 523-529.

Comment: Figures Figure 1: Consider re-moving statistical tests as these are presented in the results.

Reply: Thanks for your helpful comment. We have moved the statistical tests to the results section as suggested, please see lines 129-133, 144-147.

Reviewer #3:

Overall a very nice paper, detailed comments to the authors:

Comment: Line 35: 45X coverage is misleading since the individual coverage was much smaller, please make a clearer statement here

Reply: Thank you for this helpful suggestion. We have revised the population coverage information to individual information, please see line 36.

Comment: L40: Our FST analysis also indicates for the first time ...

Reply: Thanks for this suggestion. We have revised our manuscript according to your suggestion, please see lines 41-43.

Comment: L52: of particular economic importance ...

Reply: Many thanks for your comment. Done! Please see line 65.

Comment: L60-72: This is not introduction, but actually another summary, which I think is obsolete, a slightly more extended real introduction discussing backgraound prior knowledge, and aims of the study, would be preferred

Reply: Many thanks. We have moved this section of our previous version to Data Description according to GigaScience author guidelines and your suggestions, please lines 91-109. Meanwhile, we have revised our Introduction section, please see lines 52-85.

Comment: Figure 1B: this panel is nice, but not very informative, what exact information is retrieved from the graph?

Reply: Apologies for any problems with our figures. The complicated circos plot is the result of the many scaffolds (78,488) in the current duck reference genome. We have removed the circos plot from our current manuscript, please see Figure 1.

Comment: L95: The number of deletions was higher than the number of insertions in all nine populations

Reply: Done! Please see line 134.

Comment: L105: Move the sentence "Single base-pair INDELs were the predominant form, accounting for 38.63% of all detected INDELs (Supplemental Table S3)." before the sentence "Both the number of SNPs ..."

Reply: Thank you so much for your kind suggestion. We revised our manuscript accordingly, please see lines 142-143.

Comment: L111: ... clustered together, the three ...

Reply: Done! Please see line 155.

Comment: L117: Show figure for K=2?

Reply: Thanks for your question. Both K=2 and K=3 were showed in fig 2C, please see line 166.

Comment: L155: ... had the lowest Akaike Information Criteria (AIC) value, ...

Reply: Done! Please see lines 200-201.

Comment: L166: ... are lower than in wild mallards ...

Reply: Done! Please see line 213.

Comment: Table 1: is it possible to report standard errors or confidence intervals of the reported estimates?

Reply: Many thanks for your question. To answer the reviewer’s question we added 95% confidence intervals to all estimates. We reanalyzed the demographic history of duck domestication based on mutation rates of both zebra finch and chicken. Using the mutation rate of zebra finch (Jarvis et al. 2014), the time of duck domestication is estimated at 2,128 (+- 421) years ago. With estimates of mutation rate from chicken (Nam et al. 2010), we estimate domestication 2,228 (+- 441) years ago. Considering the genetic relationship of duck to chicken is much closer than to zebra finch (Jarvis, E. D., et al. 2014), we revised the PSMC and δaδi results of our current manuscript, please see Fig 2D, Table 1, and lines 203-211, 547-549.

Comment: L197: ... white plumage phenotype suggesting a causative mutation. Our result indicates for the first time the duck white plumage associated with selection at ...

Reply: Done! Please see lines 245-247.

Comment: L213: of 10kb size.

Reply: Done! Please see line 267.

Comment: L224: "... scaffolds longer than 10-kb by 10-kb windows with 5-kb steps." This is not clear to me, please describe better.

Reply: Apologies for any confusion. In our study, both FST and π were calculated for each 10kb size window, with 5kb size steps. However, of the 78,488 scaffolds in the duck reference genome, there are many scaffolds < 10kb. These short scaffolds were removed, and we only calculated FST for scaffolds > 10kb. We have added this to our revised manuscript, please see lines 279-281.

Comment: L237 was shown

Reply: Done! Please see lines 293-294.

Comment: L240 level differs between domesticated and wild duck.

Reply: Done! Please see line 296.

Comment: L245 I understand that you limited the GO analysis to certain processes, what happened if you included other processes as well?

Reply: Many thanks for this suggestion. In this study, all 292 genes located in the 5% FST regions (supplementary table S5) were used for the GO analysis, resulting in a total of 57 GO enrichment terms, which have now all been added to our current manuscript, please see lines 300-301, and supplementary table S6. This high number of GO terms presents a hopelessly difficult and complicated analyses, therefore we selected a subset of GO terms for further analysis based on P-value (supplementary table S7) combined the phenotypic differences between wild mallard and domestic duck. We do agree with the reviewer that a more inclusive analysis would be preferable, but the large number of GO terms makes it impossible to obtain meaningful results.

Comment: L252 identified as being under positive selection

Reply: Corrected! Please see line 311.

Comment: L258 Is "neuronal genes" the right term?

Reply: Apologies for any confusion. “Neuronal genes” is not in fact a GO term, rather a simplification of “25 neuro-synapse-axon genes” in line 310. To be more understandable, we have removed this simplification in our revision, please see line 317.

Comment: L260 fatty acid

Reply: Apologies and corrected! Please see line 319.

Comment: L269 and no gene in breast muscle

Reply: Done! Please see line 329.

Comment: L273 The results suggest that the PDC gene is of substantial functional importance in phenotypic differentiation among wild and domestic ducks.

Reply: Many thanks. We have revised this sentence according to your suggestion, please see lines 333-335.

Comment: L289 catalogued 36.1M SNPs and 3.1M INDELs,

Reply: Corrected! Please see line 349.

Comment: L333 ... showed particularly strong signs of selective sweep s presumably associated with domestication.

Reply: We have corrected our manuscript according to your suggestion, please see lines 398-399.

Comment: L340 brain and liver of domesticated ducks compared to ...

Reply: Corrected! Please see line 405.

Comment: L351 differential selection? Do you mean directional selection?

Reply: Apologies for any confusion. We also revised our current manuscript, please see lines 416-418.

Comment: L362 Taken together, our results show that duck domestication was a relatively recent and ...

Reply: We have corrected our manuscript according to your suggestion, please see line 430.

Comment: L440 From the 28,199,227 SNPs not confirmed by dbSNPs, 390 randomly chosen (?) nucleotide sites

Reply: Many thanks for your question. Of course, all nucleotide sites were randomly selected. We have added this explain to our current manuscript, please see lines 510-513.

Comment: L448 Principal Component Analysis (PCA), first by generating the genetic relationship matrix (GRM) from which the first 20 eigenvectors were extracted.

Reply: We have corrected our manuscript according to your suggestion, please see line 520-522.

-- Please also take a moment to check our website at http://giga.edmgr.com/l.asp?i=25723&l=YHKU51UQ for any additional comments that were saved as attachments. Please note that as GigaScience has a policy of open peer review, you will be able to see the names of the reviewers.

In my opinion, this revision adequately answers most of my comments. The manuscript has also improved with the answers to the other reviewer.x000Dx000D I have only a few remaining comments. The most serious one is about data availability and protocols.x000Dx000D The revision comes with better data availability. VCF files of variants are included, plus a couple of perl scripts used to process them. However, full population genetic statistics and sweep locations still seem to be missing. Scripts for running the bioinformatic tools are not included. The description of the PCR follow-up of variants has been expanded. However, the description does not include the full protocol, and neither does the description of any of the other laboratory methods. This level of detail is about the standard in the field, but it does not seem to live up to the policies of the journal.x000Dx000D A couple of times (the justification for the mix of sequence coverages, and the detail about the origin of the ducks), the reply to reviewers contain useful information that was not incorporated in the manuscript. In my opinion, the Methods should include this information, and in particular as much detail as possible about the origin of the animals.x000Dx000Dx000Dx000D Minor comments_x000D_ x000D The reply to reviewers describe the variant filtering as "extremely strict". In fact, it seems to be mostly the default starting criteria suggested by GATK developers in their "best practices" (with a "QUAL" cutoff and a higher "QD" cutoff). How were these filter settings chosen? Are they actually "extremely strict"?x000Dx000D Line 247: What does "completely associated with selection" mean in this context?x000Dx000D Lines 252-253: In what sense did the PCR primer design fail? Were you unable to amplify the region, amplify specifically, or unable to find primers that lived up to your quality criteria? I fully understand that PCR primer design fails occasionally, but I think a more specific description would be useful.x000D

Are the methods appropriate to the aims of the study, are they well described, and are necessary controls included? If not, please specify what is required in your comments to the authors.
Yes

Are the conclusions adequately supported by the data shown? If not, please explain in your comments to the authors. Yes

Does the manuscript adhere to the journal’s guidelines on minimum standards of reporting? If not, please specify what is required in your comments to the author
No

Are you able to assess all statistics in the manuscript, including the appropriateness of statistical tests used? (If an additional statistical review is recommended, please specify what aspects require further assessment in your comments to the editors.)
Yes, and I have assessed the statistics in my report.

Quality of written English Please indicate the quality of language in the manuscript:
Acceptable

Declaration of competing interests Please complete a declaration of competing interests, consider the following questions: Have you in the past five years received reimbursements, fees, funding, or salary from an organization that may in any way gain or lose financially from the publication of this manuscript, either now or in the future? Do you hold any stocks or shares in an organization that may in any way gain or lose financially from the publication of this manuscript, either now or in the future? Do you hold or are you currently applying for any patents relating to the content of the manuscript? Have you received reimbursements, fees, funding, or salary from an organization that holds or has applied for patents relating to the content of the manuscript? Do you have any other financial competing interests? Do you have any non-financial competing interests in relation to this manuscript? If you can answer no to all of the above, write ‘I declare that I have no competing interests’ below. If your reply is yes to any, please give details below.
I declare that I have no competing interests

I agree to the open peer review policy of the journal. I understand that my name will be included on my report to the authors and, if the manuscript is accepted for publication, my named report including any attachments I upload will be posted on the website along with the authors' responses. I agree for my report to be made available under an Open Access Creative Commons CC-BY license (http://creativecommons.org/licenses/by/4.0/). I understand that any comments which I do not wish to be included in my named report can be included as confidential comments to the editors, which will not be published.
I agree to the open peer review policy of the journal.

Authors' response to reviews: Reviewer #1: In my opinion, this revision adequately answers most of my comments. The manuscript has also improved with the answers to the other reviewer.

I have only a few remaining comments. The most serious one is about data availability and protocols.

Comment: The revision comes with better data availability. VCF files of variants are included, plus a couple of perl scripts used to process them. However, full population genetic statistics and sweep locations still seem to be missing. Scripts for running the bioinformatic tools are not included. The description of the PCR follow-up of variants has been expanded. However, the description does not include the full protocol, and neither does the description of any of the other laboratory methods. This level of detail is about the standard in the field, but it does not seem to live up to the policies of the journal.

Reply: Many thanks for your positive comments and apologies for any inadequate descriptions. All population genetic raw data and command scripts have been submitted to the GigaDB database. We used a sliding windows method for FST calculation in our sweep analysis, as this approach is more robust and informative for genome-wide evaluation. This approach means that one window might have several genes, and some very long genes may be present in multiple overlapping windows. Thus, we substituted sweep locations for gene locations, and added this information to our current manuscript, please see supplemental tables S5 and S8.

We have provided a citation for the specific PCR validation methods (Van et al 2008), which has been widely used in previous studies (Wang et al 2016, Yan et al 2014), please see line 536.

Van Tassell, C. P., et al. (2008). "SNP discovery and allele frequency estimation by deep sequencing of reduced representation libraries." Nat Methods 5(3): 247-252. Wang, M. S., et al. (2016). "Positive selection rather than relaxation of functional constraint drives the evolution of vision during chicken domestication." Cell Res 26(5): 556-573. Yan, Y., et al. (2014). "Genome-wide characterization of insertion and deletion variation in chicken using next generation sequencing." PLoS One 9(8): e104652.

Comment: A couple of times (the justification for the mix of sequence coverages, and the detail about the origin of the ducks), the reply to reviewers contain useful information that was not incorporated in the manuscript. In my opinion, the Methods should include this information, and in particular as much detail as possible about the origin of the animals.

Reply: Many thanks for your suggestion. We have add the justification of coverage to the Methods section of our current manuscript, please see lines 486-490. We have also detailed the point of origin for our samples, please see lines 468-474.

Minor comments

Comment: The reply to reviewers describe the variant filtering as "extremely strict". In fact, it seems to be mostly the default starting criteria suggested by GATK developers in their "best practices" (with a "QUAL" cutoff and a higher "QD" cutoff). How were these filter settings chosen? Are they actually "extremely strict"?

Reply: Many thanks for your questions. Of course, all variants were filtered with “hard filter” criteria suggested by GATK developers. However, to identify variants associated with white plumage traits, the “extremely strict” criteria were used, where variant allele frequency must be 0 in all white duck individuals and be 1 in all non-white duck individuals. Or, 1 in all white duck individuals and 0 in all non-white duck individuals. In other words, the variant had to be completely associated with the phenotype to pass our strictest threshold.

Comment: Line 247: What does "completely associated with selection" mean in this context?

Reply: Thanks for your question. “The duck white plumage is completely associated with selection at the MITF locus” means the mutations were completely associated with white plumage phenotype.

Comment: Lines 252-253: In what sense did the PCR primer design fail? Were you unable to amplify the region, amplify specifically, or unable to find primers that lived up to your quality criteria? I fully understand that PCR primer design fails occasionally, but I think a more specific description would be useful.

Reply: We were unable to design suitable primers to amplify this region, and we add this explanation to our current manuscript, please see line 270.

Reviewer #2: The revised version of the manuscript entitled, "Whole-genome resequencing reveals signatures of selection and timing of duck domestication" tackles the genomic question of domestication. The authors have done much to improve the manuscript. While most of my comments are now minor, there are a few additional requests that would be nice to see incorporated in order to strengthen the manuscript. I believe that the paper will be ready for submission if the authors incorporate all/most comments (See below).

Comment: INTRODUCTION/DATA DESCRIPTION: I think the introduction is much improved. In addition to minor comments below, I would still like to see the authors develop at least one hypothesis as to what genes/genetic regions may be playing a role in the meat/egg domestication process of these ducks. Alternatively (or in addition to), I would like to see a hypothesis regarding what they think some of the differences may be between wild and domesticated populations.

Reply: Thank you very much for your positive comments. Respectfully, the advantage of comparative genomic studies such as ours is that they are agnostic screens of the entire genome without a priori need to develop specific hypotheses. Previous similar studies of domestication (including Rubin et al. Nature 2010; Vonholdt et al. Nature 2010; Montague et al. PNAS 2014, among many others) have used these approaches to identify regions of the genome affected by artificial selection without a priori hypotheses. We adapted these approaches to the study of ducks here, with the broad aim of identifying whether ducks were domesticated once (null hypothesis) or separately for egg and meat breeds (alternative hypothesis). Moreover, we assess the role of domestication on genes related to plumage and neuroanatomy. We respectfully suggest that to develop further post hoc hypotheses to fit our results at this point would be disingenuous, and defeat the purpose of these sorts of agnostic screens.

Rubin, C. J., et al. (2010). "Whole-genome resequencing reveals loci under selection during chicken domestication." Nature 464(7288): 587-591. Vonholdt, B. M., et al. (2010). "Genome-wide SNP and haplotype analyses reveal a rich history underlying dog domestication." Nature 464(7290): 898-902. Montague, M. J., et al. (2014). "Comparative analysis of the domestic cat genome reveals genetic signatures underlying feline biology and domestication." Proceedings of the National Academy of Sciences 111(48): 17230-17235.

Comment: Line 63: remove scientific name as you already introduced mallards in the previous paragraph.

Reply: Done! Please see line 72.

Comment: Line 92: insert "of" - "….613.37 [of] Gb high….". I would also advise the authors to move any kind of findings of this type to RESULTS.

Reply: Done! Please see lines 89-91, 111-112.

Comment: Lines 94: Delete "we detected"

Reply: Done! Please see line 92.

Comment: Line 94: consider change " …,we tested for population structure between domesticated and wild populations, as well as assessed for signatures of selection associated with domestication."

Reply: Many thanks for your helpful suggestion. We have revised our manuscript accordingly, please see lines 92-96.

Comment: Line 96-98: Either delete the sentence starting with "We inferred…" or add another 1-2 sentence explaining what exactly you tested.

Reply: Deleted! Please see line 95-97.

Comment: Lines 104-109: This seems forced and out of place. Either delete it and put it to the discussion OR expand/edit it to be more streamlined.

Reply: This paragraph have been moved to discussion section, please see lines 100-105, and 449-454.

ANALYSIS:

Comment: Line 117: end with "…78 ducks."

Reply: Done! Please see line 113.

Comment: 2nd Paragraph: "Across samples, a total of 36.1 million (M) SNPs (average per sample = 4.5 M SNPs; range = 2.34 - 9.52 M SNPs) and 3.1M INDELs (average per sample = 0.4M INDELs; range = 0.21 - 0.89M INDELs) were detected (Fig. 1C1B, Supplemental Figs. S1-S2, Supplemental Table S2). ingle base-pair INDELs were the predominant form, and accounting for 38.63% of all detected INDELs (Supplemental Table S3). Our dataset covers 96.2% of the duck dbSNP database deposited in the Genome Variation Map (GVM) (http://bigd.big.ac.cn/gvm/)." In general, domesticated stock showed lower number of SNPs (t test, p = 3.13 × 10−12) and nucleotide diversity (ttest, p = 2.20 × 10−16) as compared to wild mallards (Fig. 1B - C). Moreover, homozygousity in domesticated ducks was significantly higher than ratios in wild mallards (t test, p = 1.35 × 10−10 ) consistent with the larger panmictic wild population.

Reply: Thank you so much for your helpful suggestion. This paragraph was revised accordingly, please see lines 126-151.

Comment: Line 137: does 36.1 million SNPs include indels? If not, I would just include the 2 in one summation of total diversity.

Reply: Many thanks for your question and helpful suggestion, the 36.1 million SNPs did not include INDELs. These two variation types are summed together according to your suggestion in our current manuscript, please see line 127.

Comment: Line 142 - 143: The sentence "Single base-pair INDELs were the predominant form, accounting for 38.63% of all detected INDELs (Supplemental Table S3)."

Reply: Revised! Please see lines 131-132.

Comment: Line 148: Are you sure that your data is "consistent with larger panmictic wild population" ? What about artificial selection and inbreeding within domesticated stock? Maybe both? Consider revising.

Reply: Apologies for any confusion. We had revised our manuscript accordingly, please see lines 138-140.

Comment: Lines 155 - 158: Consider changing the sentence to: "In general, clustering among samples corresponded with their source, that included wild ducks (MDN and MDZ), ducks domesticated for meat production (PK, CV, and ML), and ducks domesticated for egg production (JD, 157 SM, and SX). The dual-purpose domesticate clustered with ducks domesticated for egg production (Fig. 2B-C)."

Reply: Done! Please see lines 156-160.

Comment: Lines 184-202: Consider revising to 1 paragraph: "Next, we explored the demographic history of our samples to differentiate whether domestication of meat and egg producing ducks was the result of one or multiple events. First, we estimated changes in effective population size (Ne) in our three genetic clusters in a pairwise sequentially Markovian coalescent (PSMC) framework [22]. The meat type ducks (PK, CV, and ML) showed concordant demographic trajectories with egg and mixture dual-purpose type populations (JD, SM, SX, and GY) with one apparent expansion around the Penultimate Glaciation Period (PGP, 0.30-0.13 Mya) [4, 23] and Last Glacial Period (LGP, 110-12 kya) [24, 25], followed by a subsequent contraction (Fig. 2D). Next, we tested multiple demographic scenarios …."

Reply: Done! Please see lines 187-208.

Comment: Line 214: What is the Ne for the wild population. Please make clear by at least referencing Table 1.

Reply: Thank you for this helpful suggestion. We have had add the Ne estimate of the wild population to our main text, please see line 225.

Comment: Lines 224-229: Please cite sources for some of your statements here. Better to make the statement of your findings and save lines 226-229 for discussion.

Reply: Many thanks for your comments. We have moved lines 226-229 to the discussion section according as suggested, please see lines 387-390.

Comment: Line 241: I would like to know if any other region showed deviation/outliers? Or was there only 1 region across the entire genome? Please clarify.

Reply: Many thanks for your questions. This region is the fourth ranked region across the entire genome, but the only one region correlated with coloration. We also revised our current manuscript, please see lines 251-261.

Comment: DISCUSSION: Overall, the discussion is well written, organized, and I find the topics of broad appeal.

I believe the introduction of the Discussion can be combined into a single paragraph and a bit streamlined as it is just reiterating the results.

Reply: Thank you so much for your positive comment and your helpful suggestion. The introduction of the Discussion have been revised and redundant material deleted as you suggest, please see lines 348-363.

Comment: Lines 348 - 353: Consider splitting into at least 2 sentences.

Reply: Done! Please see lines 357-363.

Comment: Line 419: add "and": "dogs [45], and…"

Reply: Done! Please see line 433.