The Value of Reference Genomes in the Conservation of Threatened Species

Round 1

Reviewer 1 Report

The authors present a manuscript aimed at highlighting the value of a reference genome for conservation using the case study of the Tasmanian devil. The manuscript was very well-written and was a genuine pleasure to read. I believe it can make a strong contribution to the conservation genomics literature.

A couple of general points:

First, a key goal of the manuscript is to highlight how a reference genome is useful for conservation. It did a good job of highlight the potential uses for a reference genome in terms of the kinds of data/information that could be generated, but I felt the authors could have provided some more detail about how the reference genome has actually informed on-the-ground conservation. It no doubt has! – I just felt the authors did not provide enough context to readers who are not familiar with the system and/or genomics for this to come across clearly. Whilst the value of a reference genome is alluded to often, a section dedicated to highlighting some of the specific conservation actions/decisions that were informed by genomic information and the subsequent conservation outcomes was something critical I felt was missing from this manuscript. Essentially what I would have liked to see is a section that explicitly highlights how conservation outcomes have been improved for the Tasmanian devil as a result of using genomic information to inform management. I wondered if some sort of table could also be helpful here – e.g. it could have the different conservation decisions actions, outcomes, and how each specific decision/action was assisted by having access to a reference genome.

Second, I thought it might be beneficial to include some ball-park costs currently associated with development of a high-quality reference genome (e.g. what was the approximate cost of the Tasmanian devil genome?). I understand that the field is rapidly progressing and so current costs may not stay current for long, but I still feel it could be helpful information to include, especially if the authors want to make the argument that a reference genome is an affordable option for conservation programs. This would help readers gauge the feasibility of developing a reference genome for their own system – a reference genome of high-enough quality to be useful is arguably still unaffordable for many working in conservation.

Specific comments:

Line 11: “few are taking advantage of the full range of genomic technologies available today at affordable prices.” I would argue that whilst prices have fallen dramatically, “affordable” is still going to be context dependent – genomic methods are still going to be too expensive for many working in conservation with small budgets. I feel it is important for this paper to be sensitive to the whole spectrum of conservation (the Tasmanian devil conservation program is a program that receives a large amount of funding relative to many other less iconic plants and animals). Consider rephrasing this bit.

Line 42: “Maintaining genetic diversity is an important component of population viability as it arms populations with the potential to adapt to future environmental change” – I would also argue not just for future adaptive potential but also for immediate population viability, due to costs of inbreeding depression.

Line 78: “Targeted education and training is also required to teach conservation managers how to interpret and utilise genomic data.” This seems non-trivial. Should this really be an expectation of conservation managers?

Line 156: “ensuring they were representative of genome diversity” – can 22 microsatellites really be representative of genome diversity? This is still a pretty small number of loci (representing one kind of mutation pattern). Perhaps rephrase.

Line 227-233 – It could be good to expand either here or elsewhere some of the detail around the examples of how a reference genome benefitted devil conservation given here. For example, what decisions were made based on “genetic variation analyses” and what were the conservation outcomes? How did the genome help with “selection of individuals for release into the wild” and what were the conservation outcomes? etc

Line 271: “as of 2017, WGR routinely costs over $1000 per individual whereas RRS costs less than $100 272 per individual” - can the authors provide a more up-to-date cost estimate?

Table 1. This table did not seem all that essential to me – does it really need to be included? The authors seem to imply that the current minimum reference genome quality standards are potentially problematic and imposing barriers for conservation (e.g. Line 327), but do not really make any recommendations based on this.

Line 401: “Collaborating with global genome consortia (like the Earth Biogenome Project) is the best way for independent conservation programs to utilise the full potential of genomic resources and join the genomics revolution.” It might be good to expand on this statement earlier on in the manuscript by providing more explicit discussion of the role of genome consortia – Is the development of reference genomes best left to consortia? Is it inefficient for researchers to pursue developing reference genomes independently? Are consortia always obliged to make their genomes publicly available?

Author Response

First, a key goal of the manuscript is to highlight how a reference genome is useful for conservation. It did a good job of highlight the potential uses for a reference genome in terms of the kinds of data/information that could be generated, but I felt the authors could have provided some more detail about how the reference genome has actually informed on-the-ground conservation. It no doubt has! – I just felt the authors did not provide enough context to readers who are not familiar with the system and/or genomics for this to come across clearly. Whilst the value of a reference genome is alluded to often, a section dedicated to highlighting some of the specific conservation actions/decisions that were informed by genomic information and the subsequent conservation outcomes was something critical I felt was missing from this manuscript. Essentially what I would have liked to see is a section that explicitly highlights how conservation outcomes have been improved for the Tasmanian devil as a result of using genomic information to inform management. I wondered if some sort of table could also be helpful here – e.g. it could have the different conservation decisions actions, outcomes, and how each specific decision/action was assisted by having access to a reference genome.

We have added a table (Table 1) to present how the reference genome has been used to inform conservation based on the reviewer’s recommendation and have referred to this table throughout the manuscript.

Second, I thought it might be beneficial to include some ball-park costs currently associated with development of a high-quality reference genome (e.g. what was the approximate cost of the Tasmanian devil genome?). I understand that the field is rapidly progressing and so current costs may not stay current for long, but I still feel it could be helpful information to include, especially if the authors want to make the argument that a reference genome is an affordable option for conservation programs. This would help readers gauge the feasibility of developing a reference genome for their own system – a reference genome of high-enough quality to be useful is arguably still unaffordable for many working in conservation.

We have included a rough cost estimate of reference genome creation at lines 67-70. “Creating high-quality reference genomes that can provide insights into species evolution and biology is a costly task (~$30,000 for an average eukaryotic genome size of 2.5Gbp [20]), and also requires large collaborative groups to provide expertise from varying fields (e.g. [21-23]).”  We appreciate that some groups are likely to lack the resources/funding to develop high-quality reference genomes themselves, so have included the following information (lines 436-441) to clarify that our recommendation in these situations is to collaborate with global genome consortia (or more local sequencing efforts). “We would recommend that conservation managers who are seeking to use the types of methods we have described herein collaborate with global genome consortia (like the Earth Biogenome Project) or national/local consortia (like the Oz Mammal Genome Initative) to utilise the full potential of genomic resources and join the genomics revolution. This allows conservation managers to focus on conservation and work with geneticists who can help them make adaptive management decisions in real time [37].”

Specific comments:

Line 11: “few are taking advantage of the full range of genomic technologies available today at affordable prices.” I would argue that whilst prices have fallen dramatically, “affordable” is still going to be context dependent – genomic methods are still going to be too expensive for many working in conservation with small budgets. I feel it is important for this paper to be sensitive to the whole spectrum of conservation (the Tasmanian devil conservation program is a program that receives a large amount of funding relative to many other less iconic plants and animals). Consider rephrasing this bit.

This is an important consideration, so we have removed “at affordable prices” from the abstract and have also rephrased a similar sentence in the final paragraph (lines 431-434) to “Reduced costs and lower input DNA requirements, as well as improved bioinformatic assembly and annotation pipelines based on non-model non-eutherian species, mean that these technologies are becoming more attainable by conservation programs and should be used more routinely where budgets allow [95].”

Line 42: “Maintaining genetic diversity is an important component of population viability as it arms populations with the potential to adapt to future environmental change” – I would also argue not just for future adaptive potential but also for immediate population viability, due to costs of inbreeding depression.

We have rephrased the sentence to “Maintaining genetic diversity is an important component of population viability as it assists with mitigating negative effects associated with inbreeding and arms populations with the potential to adapt to future environmental change”.

Line 78: “Targeted education and training is also required to teach conservation managers how to interpret and utilise genomic data.” This seems non-trivial. Should this really be an expectation of conservation managers?

We have added the following additional information and associated references to this section of the manuscript (lines 83-92) to better explain ways in which genetic research can be better implemented into conservation management. “To better assist conservation managers, a number of groups and communities have already been established to assist in providing conservation genetics advice for threated species management. These include the IUCN/SSC (Species Survival Commission) Conservation Genetics Specialist Group (CGSG), the Genetic Composition Working Group of GEO BON (Group on Earth Observations Biodiversity Observation Network) and the pan-European COST (Cooperation in Science and Technology) action ConGRESS (Conservation Genetic Resources for Effective Species Survival) (for further information and examples from these groups see Holderegger, et al. [34]). Conservationists in their respective countries can get in touch with these groups to obtain the contact details of geneticists who work in their region who may be able to assist them with their management needs.”  

Line 156: “ensuring they were representative of genome diversity” – can 22 microsatellites really be representative of genome diversity? This is still a pretty small number of loci (representing one kind of mutation pattern). Perhaps rephrase.

This has been rephrased to “providing a greater representation of neutral genome-wide diversity in comparison to the original 11 putatively neutral microsatellites”.

Line 227-233 – It could be good to expand either here or elsewhere some of the detail around the examples of how a reference genome benefitted devil conservation given here. For example, what decisions were made based on “genetic variation analyses” and what were the conservation outcomes? How did the genome help with “selection of individuals for release into the wild” and what were the conservation outcomes? Etc

We have added Table 1 which includes more detailed information on how a reference genome has benefitted Tasmanian devil conservation.

Line 271: “as of 2017, WGR routinely costs over $1000 per individual whereas RRS costs less than $100 per individual” - can the authors provide a more up-to-date cost estimate?

These figures still hold true at the present time, the main difference is that today you are able to achieve a slightly greater read depth for these prices providing even greater resolution. As a result, we have changed the year to 2019 in the manuscript.

Table 1. This table did not seem all that essential to me – does it really need to be included? The authors seem to imply that the current minimum reference genome quality standards are potentially problematic and imposing barriers for conservation (e.g. Line 327), but do not really make any recommendations based on this.

We believe this table is useful to those familiar with common genome statistics to demonstrate the potential of genomes (such as the Tasmanian devil) which are far from recently developed consortia standards. However, we understand it may not be necessary in the main manuscript, so have moved this table to supplementary materials.

Line 401: “Collaborating with global genome consortia (like the Earth Biogenome Project) is the best way for independent conservation programs to utilise the full potential of genomic resources and join the genomics revolution.” It might be good to expand on this statement earlier on in the manuscript by providing more explicit discussion of the role of genome consortia – Is the development of reference genomes best left to consortia? Is it inefficient for researchers to pursue developing reference genomes independently? Are consortia always obliged to make their genomes publicly available?

We have added some additional information in the introduction where we first mention consortia (lines 76-79) to clarify that the main role of genome consortia is to bring together the required expertise to generate reference genomes of a sufficient quality for the science community. While consortia are not always obliged to make their genomes publicly available, that is often a high priority and we have mentioned this in the following sentence. “The goal of many of these consortia is to bring together the required expertise to generate reference genomes of a sufficient quality which are publicly available to the science community, thereby providing the vital resources required to better implement genomics into conservation management [13,15,18].” We have also described why the creation of reference genomes is not often feasible for researchers to pursue independently in lines 67-70 “Creating high-quality reference genomes that can provide insights into species evolution and biology is a costly task (~$30,000 for an average eukaryotic genome size of 2.5Gbp [20]), and also requires large collaborative groups to provide expertise from varying fields (e.g. [21-23]).” and have more clearly stated in our final conclusions that our recommendation is to collaborate with consortia where possible (lines 436-441) “We would recommend that conservation managers who are seeking to use the types of methods we have described herein collaborate with global genome consortia (like the Earth Biogenome Project) or national/local consortia (like the Oz Mammal Genome Initative) to utilise the full potential of genomic resources and join the genomics revolution. This allows conservation managers to focus on conservation and work with geneticists who can help them make adaptive management decisions in real time [37].”

Author Response File: Author Response.docx

Reviewer 2 Report

Summary

This review summarizes the benefits of having a reference genome for conservation initiatives of endangered species, using the Tasmanian Devil as a case study. The authors discuss how the availability of a reference genome enabled researchers to develop additional genetic tools like microsatellites and SNP panels relatively efficiently and inexpensively. These tools have then been used to genotype wild individuals non-invasively from scat samples, generate a pedigree of the insurance population, improve estimates of inbreeding and genome-wide heterozygosity, and assess mate choice in captivity, and identify genes under selection that might improve treatment of Devil Facial Tumor Disease. This review is well-written and timely given the extreme loss of biodiversity in the current age

 

Broad Comments

This review makes an extremely compelling argument for the utility of a reference genome for the Tasmanian Devil, but much of the reasons presented by the authors are due to the relatively unique situation of the Tasmanian Devil, in terms of the decline of the Tasmanian Devil due to DFTD. Although many amphibians have suffered similar declines due to chytridiomycosis, most other species are threatened by habitat loss, climate change, invasive species, pollution, and exploitation via hunting and the wildlife trade. Thus, the authors need to broaden the review somewhat in order to increase the applicability of this case study. For example, in lines 105-109, although many threatened species have low genome-wide diversity, is it so low that microsatellites cannot provide an accurate picture of founder relatedness and population substructure? Also, how does this apply to species with moderate or high genome-wide diversity? Also, this application is useful for primer generation of any candidate gene of interest, greatly increasing the ease of investigative genetic work. After each discussion of a particular benefit, I would like to see a few sentences about how this type of work might apply more widely to other, more traditional endangered species.

The authors mention in lines 59-64 that wildlife and conservation management teams may lack the expertise to develop and utilize genomic data. Additionally, lines 78-79 point out, “Targeted education and training is also 78 required to teach conservation managers how to interpret and utilise genomic data.” This review would be improved by the mention of some tools or educational training that managers could benefit from.

The costs of whole-genome resequencing versus reduced representation sequencing are mentioned – it might also be useful to mention the cost of developing the SNP panel or developing X microsatellites with a genome versus without.

In the paragraph beginning in line 277, the authors discuss performing a genome-wide association study, without discussing the caveats that for most species, 10 individuals will not be sufficient to identify underlying genetic basis of a phenotype. The selective pressure of DFTD is somewhat unique, meaning that fewer individuals are needed to identify genes that contribute to resistance. Some discussion this trade-off should be added.

In lines 368, please elaborate briefly on the limitations on the current fragmented reference assembly.

 

 

Specific Comments

When referencing the published genomes in lines 111-114, 121-122, and 339-340, why are so many authors mentioned? Please edit.

Author Response

This review makes an extremely compelling argument for the utility of a reference genome for the Tasmanian Devil, but much of the reasons presented by the authors are due to the relatively unique situation of the Tasmanian Devil, in terms of the decline of the Tasmanian Devil due to DFTD. Although many amphibians have suffered similar declines due to chytridiomycosis, most other species are threatened by habitat loss, climate change, invasive species, pollution, and exploitation via hunting and the wildlife trade. Thus, the authors need to broaden the review somewhat in order to increase the applicability of this case study. For example, in lines 105-109, although many threatened species have low genome-wide diversity, is it so low that microsatellites cannot provide an accurate picture of founder relatedness and population substructure? Also, how does this apply to species with moderate or high genome-wide diversity? Also, this application is useful for primer generation of any candidate gene of interest, greatly increasing the ease of investigative genetic work. After each discussion of a particular benefit, I would like to see a few sentences about how this type of work might apply more widely to other, more traditional endangered species.

We have broadened throughout the manuscript on how this can be applied to other threatened species. This is done particularly at:

Lines 104-106 “Although this species has a unique conservation issue, low genetic diversity coupled with an infectious clonal cancer, the methods described herein apply to many other threatened species.”

Lines 127-130 “In instances such as these, further genomic data was required to improve resolution. For other threatened species, where there may be moderate to high genome-wide diversity, microsatellite markers may be highly polymorphic and so these markers have value as a continuing genetic management tool.”

Lines 234-238 “Recent work investigating New Zealand threatened bird species also showed the benefits of calling SNPs against conordinal, confamilial, cogeneric and conspecific reference genomes [95]. This highlights that not every threatened species requires a reference genome, although the quality of the SNP data reduces as you move away from the genus and family level.”

Lines 420-422 “Tasmanian devils are not the only species who are threatened globally by disease, other examples include black-footed ferret and distemper [120], bats and white-nose syndrome [121], and frogs and chytrid [122].”

 

To summarise the applications we have added the following information at the end of the review (lines 442-453). “Although here we have presented a unique case study of a species with significantly low levels of genetic diversity and a large threatening disease process, the techniques described for the Tasmanian devil can be applied more broadly to many species of conservation concern. The applications of what we have described herein for devils is not unique to this species as many of the questions we have answered are posed by those managing other threatened species. These include understanding historical demography and current population structure, minimising inbreeding, maximising adaptive potential, and identifying the basis of important phenotypic traits (whether these be related to disease, behaviour or reproduction). Hence, despite differences in threatening processes and current state of vulnerable species, the nature of their small population sizes will result in a number of common conservation concerns that can be informed using genomic data [15,18]. In the midst of the sixth mass extinction event, we advocate the use of reference genomes and associated genetic tools to arm conservation managers with ways to assist the long-term survival of species.”

The authors mention in lines 59-64 that wildlife and conservation management teams may lack the expertise to develop and utilize genomic data. Additionally, lines 78-79 point out, “Targeted education and training is also 78 required to teach conservation managers how to interpret and utilise genomic data.” This review would be improved by the mention of some tools or educational training that managers could benefit from.

We have added the following information and associated references to this section of the manuscript (lines 83-92) to better explain ways in which genetic research can be better implemented into conservation management.  “To better assist conservation managers, a number of groups and communities have already been established to assist in providing conservation genetics advice for threated species management. These include the IUCN/SSC (Species Survival Commission) Conservation Genetics Specialist Group (CGSG), the Genetic Composition Working Group of GEO BON (Group on Earth Observations Biodiversity Observation Network) and the pan-European COST (Cooperation in Science and Technology) action ConGRESS (Conservation Genetic Resources for Effective Species Survival) (for further information and examples from these groups see Holderegger, et al. [34]). Conservationists in their respective countries can get in touch with these groups to obtain the contact details of geneticists who work in their region who may be able to assist them with their management needs.”

The costs of whole-genome resequencing versus reduced representation sequencing are mentioned – it might also be useful to mention the cost of developing the SNP panel or developing X microsatellites with a genome versus without.

We have added some rough cost estimates for developing microsatellite markers both with and without a genome at lines 176-180 as microsatellites are still commonly used in conservation practice. “It has previously been estimated that development of just 10 microsatellite markers without prior genetic data can cost up to $10,000 [51]. The availability of a reference genome mitigates the need for traditional microsatellite isolation procedures and therefore significantly reduces costs associated with marker development (<$1000 for primer optimisation and testing, unpublished data).” Since the cost of SNP panel development relies heavily on a variety of differing factors including the approach (amplicon sequencing vs targeted panels vs target enrichment), the number of loci, the number of samples and the required coverage etc, it is very difficult to provide a useful estimate and hence we have not included a cost for this in the manuscript.

In the paragraph beginning in line 277, the authors discuss performing a genome-wide association study, without discussing the caveats that for most species, 10 individuals will not be sufficient to identify underlying genetic basis of a phenotype. The selective pressure of DFTD is somewhat unique, meaning that fewer individuals are needed to identify genes that contribute to resistance. Some discussion this trade-off should be added.

Thank you for highlighting this is an important point, we have the following to discuss this trade-off at lines 329-334. “It is important to note that often larger numbers of individuals are required to identify genes underlying certain phenotypes, particularly in species with higher genetic diversity and/or reduced selective pressure on the phenotype of interest [111]. This requires careful consideration of trade-offs between the sequencing approach (targeted vs RRS vs WGR), number of samples and sequencing coverage, and will often depend upon some prior knowledge (or preliminary testing), budget, and access to samples.”

In lines 368, please elaborate briefly on the limitations on the current fragmented reference assembly.

We have included some examples of the major limitations of fragmented assemblies to this sentence (lines 407-409) “……. reference genomes will become easier to create and will overcome many of the limitations of current fragmented reference assemblies such as incomplete gene characterisation, comparative evolutionary limitations, and increased computational requirements. [117].”

 

Specific Comments

When referencing the published genomes in lines 111-114, 121-122, and 339-340, why are so many authors mentioned? Please edit.

This has now been amended.

 

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

The authors have addressed all of my comments in a satisfactory manner. Although still fairly narrow in scope, this paper is ready for publication.