이 장은 원문만 있습니다. 문장별 해석은 앞 장부터 차례로 붙입니다.
1장 · Conservation of biodiversity in the genomics era
“Conservation genomics” encompasses the idea that genome-scale data will improve the capacity of resource managers to protect species. Although genetic approaches have long been used in conservation research, it has only recently become tractable to generate genome-wide data at a scale that is useful for conservation. In this Review, we discuss how genome-scale data can inform species delineation in the face of admixture, facilitate evolution through the identification of adaptive alleles, and enhance evolutionary rescue based on genomic patterns of inbreeding. As genomic approaches become more widely adopted in conservation, we expect that they will have a positive impact on management and policy decisions.
Introduction
The human footprint on our planet is currently threatening biological diversity across habitats. Arguably the biggest threat to biodiversity across the planet is habitat degradation. As the human population increases, we modify the landscape to meet our increasing need for resources to support modern lifestyles. Coincident with this is an increase in energy consumption that is driving climate change across the globe. The rapid pace of the changing climate will outpace the natural ability of some species to respond. Temporal analysis of biodiversity loss indicates that we are on a trajectory for the Earth’s sixth mass extinction event, with the rate of extinction in the last century conservatively estimated to be 22 times faster than the historical baseline rate. The picture is even more bleak when the analysis examines population declines, rather than the complete loss of species, with 32% of known vertebrate species showing substantial population declines.
Efforts to stop mass extinctions and population declines include setting up protected areas (for example, marine protected areas (MPAs)), international agreements to limit greenhouse gases to curb climate change (for example, the Kyoto Protocol and the Paris Agreement), and legal frameworks to protect endangered species (for example, the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES) and the US Endangered Species Act (ESA)). Genomic technologies can aid these efforts by identifying biodiversity “hotspots” to prioritize for protection, using predictive models to help build natural communities that are resilient to environmental change, and informing management actions that attempt to mitigate threats to endangered species.
In this Review, we differentiate genetic approaches, which use a small number of neutral markers, from genomic approaches, which use complete genomes or genome-wide data. No standardized amount of data divides genetics from genomics; rather, this is a semantic distinction. We consider a study to have transitioned into the realm of genomics when a high density of markers is assayed from across the entire genome, usually in the order of thousands of markers.
Although both genetic and genomic data sets can be used to estimate genetic diversity, population structure, and demographic history, genome-scale data, with an increased density of markers across the genome, can provide more accurate estimations of these parameters, sometimes resulting in different conservation recommendations. For example, an analysis of more than 25,000 loci in the foothill yellow-legged frog revealed strong differentiation between five phylogenetic clades that the researchers suggested should provide the foundation for the management of the species; whereas a previous analysis of 1,525 bp of mitochondrial DNA (mtDNA) did not have the resolution to recover these clades and instead recommended conservation based on hydrologic boundaries. Similarly, an analysis of 3,095 single nucleotide polymorphisms (SNPs) in the eastern tiger salamander found that roads restricted movement between ponds; however, a prior study using microsatellite loci to examine the same ponds found high migration rates between ponds. The most recent study suggested that mitigation of the impact of roads on the connectivity between ponds would be an important conservation target.
In addition to the increased precision of estimates of traditional parameters, the transition to genomic approaches allows researchers to ask qualitatively different questions. This is because our capacity to examine different evolutionary mechanisms increases with the amount of the genome interrogated. In addition to assaying putatively neutral loci and protein-coding regions of the genome, whole-genome sequencing allows the identification of non-coding regulatory regions that control gene expression, and whole-transcriptome sequencing allows the quantification of gene expression differences.
The limited use of genome-scale data in a conservation context is probably due to the additional challenges presented by these data sets. One important consideration is cost. Although the cost of sequencing continues to decrease, most conservation projects have limited budgets that allow genome-scale sequencing of only a small number of samples. The tradeoff between the number of samples and the number of loci sequenced is a critical consideration, and the best approach in each case will depend on the specific research question. Another important consideration is data analysis; that is, the specific resources and expertise that are available to analyze whole-genome data. Calling genotypes requires a reference genome, which may not be available for many non-model organisms, and analysis software is not always user-friendly. Finally, once a researcher obtains results from whole-genome analyses, it is often difficult to interpret the results and to translate them into conservation recommendations.
In this Review, we discuss how conservation researchers and managers can use the power of genomic data to make decisions on the conservation of biodiversity. We focus on conservation topics where genome-scale data can provide valuable insights that are unattainable with traditional genetic techniques: delineating species in the face of admixture, identifying adaptive alleles through association mapping, and enhancing evolutionary rescue based on genomic patterns of inbreeding.
Admixture and species delineation
The current conservation regulatory framework relies on defining distinct units of conservation to support law enforcement and to inform resource allocation. In conservation, the term “species” is often used to convey the idea of a unit of conservation and includes taxonomic levels below species, such as subspecies and distinct populations. Defining specific species is fraught with challenges and differing opinions. In conservation, researchers tend to prefer a phylogenetic species concept, which identifies species based on their apparent differences, but which may over-split groups. Other common species concepts require estimations of genetic distances or proof of reproductive isolation, which are challenging data to gather from most natural populations. Disagreement over how species should be defined highlights both the artificial nature of species as purely discrete units and the importance of defining species in biology, where a means to categorize organisms provides a framework for hypothesis testing. Several discussions of the relevance to conservation of defining species have recently been published, and we refer the interested reader to these.
Identifying and describing conservation units is often confounded by the lack of clear boundaries between management units. The most common categorization currently in use is the evolutionarily significant unit (ESU), which defines a group as distinct if it is “substantially reproductively isolated from other conspecific population units” and “represents an important component in the evolutionary legacy of the species”. An ESU of a vertebrate species can be defined as a distinct population segment (DPS), which is the smallest biological categorization that can be listed under the ESA. ESUs and DPSs are populations that may be geographically isolated or that may be morphologically distinct from other populations of the same species, and may also be distinct based on some measure of genetic divergence. There is no strict rule, however, regarding the amount of genetic divergence required to qualify as an ESU or a DPS; definitions of genetic distinctiveness for ESUs range from significant divergences in allele frequencies to a consistent phylogenetic signal across multiple tested genes.
In evolutionarily simple biological systems, traditional genetic techniques can delineate conservation units in a straightforward manner. However, the task is more complicated in complex evolutionary systems, such as those with a history that includes admixture and introgression. Admixture is the interbreeding between individuals from distinct groups, such as that between two related species. Introgression is the transfer of alleles from one species to another. Admixture and introgression complicate the task of delineating units of conservation because analyses of different parts of the genome can result in qualitatively different answers. This conflict can be seen in the genomes of plains bison (Fig. 1), which have a known history of admixture with cattle. An analysis of the mtDNA of a Santa Catalina herd indicated that the herd’s ancestry is 44.9% cattle; but an analysis of the autosomal DNA indicated that the herd’s ancestry is only 0.6% cattle.
Genomic research has revealed a high frequency of admixture in natural systems, ranging from great apes to bears and butterflies. For example, evidence of admixture between ancient anatomically modern humans and archaic hominins is written into the genomes of most present-day humans, who individually contain up to 7.4% ancestry from Neanderthals and Denisovans. As genome technologies and genomic resources have improved, so have the statistical methods to detect and quantify admixture. It is now possible not only to detect ancient admixture, but also to examine the genomic signatures of admixture on a fine scale. Researchers are now able to detect rare admixture events; however, these rare events may not be critical components of the evolutionary history of the species, so their identification may cloud attempts to delineate units for conservation. More importantly, high-resolution genomic data enable researchers to infer ancestry for specific regions of the genome and to estimate the timing of admixture events.
Understanding the admixture histories of natural populations is important when delineating units for conservation, as admixture erodes the genetic distinctiveness on which conservation units are based. Historically, this has led admixture to be seen in conservation as a threat to the integrity of endangered species. More recently, as genomic research has revealed its commonness in evolutionary history, admixture has come to be viewed as a potential source of new genetic variation. In this view, the influx of new genes from admixture can be seen to provide critical variation on which natural selection can act. This new variation may be vital, for example, in highly inbred populations or in populations at the edges of their ideal habitat range where rapidly changing environments may pose a considerable threat.
Given that conservation legislation is based on the identification of distinct units, it is not surprising that regulations also vary with respect to how hybrid populations should be protected. Some conservation policies favor the eradication of admixed populations, particularly if admixture has occurred because of human intervention. Even policies that do not favor eradication tend to provide few specific guidelines for categorizing admixed populations. This practice leads to policy implementation that varies from no protection to complete protection for admixed individuals.
Although genomics will not solve the problem of discrete classification in an inherently non-discrete system, genome-scale data can provide researchers and managers with a more complete understanding of the spatial and temporal dynamics of admixture in evolutionarily complex systems. Much research in this realm has taken place in naturally occurring hybrid zones where one of the two parent species is protected. In both genetic and genomic approaches, the main goal is to identify ancestry-informative markers that are capable of distinguishing the two parent-species and estimating the proportion of ancestry of the protected parent species in hybrid individuals. For example, using a genetic approach, researchers used amplified fragment length polymorphisms (AFLPs) to determine parentage in hybrid garter snakes in Wisconsin. Despite the limitations of AFLPs as genetic markers, this research provided important insights to managers. The AFLPs proved that nearly genetically pure members of the protected species occurred throughout the garter snake hybrid zone. This finding suggests that unless morphologically diagnostic characteristics are identified, protection of the endangered garter snake would only succeed if both species were protected in regions where their ranges overlapped.
As DNA-sequencing technologies advance, so does the ability to sequence markers more densely across the genome, which both improves parentage estimates and provides a means to identify patterns of genetic introgression, with potential conservation implications. For example, researchers used expressed sequence tags (ESTs) to identify SNPs that were fixed for different alleles between the threatened California tiger salamander and the intentionally introduced non-native barred tiger salamander. Researchers identified 68 ancestry-informative SNPs and used these SNPs to quantify ancestry. They then tracked the spread of these invasive alleles by mapping marker allele frequency against geographic distance from a known introduction site. Although 65 invasive alleles did not spread far from the introduction sites, the remaining three have moved 90 km in the 60 years since the introductions began, indicating that alleles can move at different rates across the landscape. Additional insights into the implications of hybridization may be obtained through whole-genome sequencing, which takes a more complete look at the genome than do ESTs by allowing the interrogation of unexpressed regions of the genome, such as non-coding regulatory sequences. However, for these salamander species whole-genome sequencing is currently impractical, as both species have genomes in excess of 30 Gb. Exome-capture methods are in development to provide high-density genome-wide markers with the aim of addressing these questions. From a conservation perspective, this research highlights how difficult it might be to contain invasive alleles once they are introduced, suggesting that a goal of maintaining “pure” species in the face of hybridization may be impractical.
Genomic inference can, however, help to protect specific traits. For example, by combining fine-scale genomic data with phenotypic data, it is possible to connect particular genomic regions to ecologically important traits (as discussed below in the “Association mapping and adaptation” section). This would provide an improved understanding of the ecological consequences of introgression and may lead to targeted efforts to protect individuals that carry these traits. The ability to identify adaptive genetic variation raises the possibility of using this variation to delineate conservation units. For example, a proposal has been submitted to list spring-run Chinook in the Klamath River as endangered under the ESA based on a single allele that is strongly associated with the spring-run phenotype. This proposal, however, has been controversial, with some parties highlighting the need to protect adaptive variation, and other parties more concerned about the implications of conservation decisions based on single-gene analyses. Others argue that, rather than focusing on a few genes and traits of interest, delineation of conservation units should include genome-wide signatures of adaptation. Important to this ongoing debate is to acknowledge that methods used to identify genome-wide adaptive variation are hampered by high rates of false negatives and false positives. The outcome of this specific debate is as-yet unknown, but will no doubt have implications for conservation genomics research and practice.
Another notable example of using genome-scale data to delineate conservation units is in the wild canids of North America. Whole-genome sequencing was used to detect admixture and to disentangle the complex evolutionary history of wolves and coyotes (Box 1). Gray wolves in North America have been divided into several subspecies: Canis lupus baileyi, Canis lupus nubilus, Canis lupus occidentalis, Canis lupus arctos, and Canis lupus lycaon (the eastern wolf) (Fig. 2). The taxonomic status of the eastern wolf has been controversial in large part due to a complex history of admixture with coyotes. This has implications for conservation because the eastern wolf is currently protected as a subspecies of gray wolf. However, the US Fish and Wildlife Service (USFWS) has suggested that the eastern wolf is instead a longstanding lineage native to eastern North America that was derived from a common ancestor with the coyote and has recently admixed with gray wolves. Given this taxonomic revision, the eastern wolf is not protected under the gray wolf ESA listing. Using a high-density domestic dog SNP chip and whole-genome sequences, researchers found that qualitative patterns of variation across the genome indicate that the eastern wolf is of gray wolf ancestry with recent admixture with coyotes. They estimated the time since admixture using SABER software, which models ancestry blocks using a Markov-hidden Markov model (MHMM) and accounts for ancestral linkage disequilibrium. They estimated that admixture occurred approximately 600–900 years ago, which is prior to the invasion of coyotes into areas occupied by the eastern wolf. They inferred that admixture with coyotes may have been an important component in eastern wolf evolution. However, SABER does not model haplotype structure, which provides additional information on the timing of admixture events. Additionally, programs such as SABER only estimate the time since admixture when there has been a single admixture event. The development of new statistical methods that can disentangle multiple admixture events that occur across the evolutionary history of a species will be informative for this and other conservation decisions relating to admixed species.
Box 1: Conservation implications of admixture in the eastern wolf
The history of federal protection of the gray wolf in the US spans 50 years. Initially, individual gray wolf subspecies were protected separately. In 1978, the US Fish and Wildlife Service (USFWS) issued a ruling under the US Endangered Species Act (ESA) to reclassify the gray wolf as endangered at the species level, protecting gray wolves throughout the lower 48 states and Mexico. Some gray wolf populations have since recovered and six states have been removed from the 1978 listing. Gray wolves are currently protected in 42 states and Mexico. In 2013, the USFWS proposed delisting the gray wolf based on a taxonomic revision by USFWS scientists. The revised taxonomy considers the eastern wolf subspecies, C. lupus lycaon, a separate species, C. lycaon, and means that the current listing for C. lupus is invalid as it includes 29 states that are occupied by C. lycaon rather than by C. lupus. In addition, in reassessing the status of C. lupus based on the new taxonomy, the USFWS found that C. lupus was neither threatened nor endangered, with the exception of the subspecies C. l. baileyi in the southwestern US and Mexico.
The 2013 taxonomic revision that led to the proposed delisting of the gray wolf has proved to be as controversial as other aspects of wolf protection and recovery in the US. Although there seems to be agreement that admixture is an important component in explaining patterns of genetic variation in eastern wolves, there is disagreement about the context of admixture and the implications for canid taxonomy. Chambers et al. (2012) argue that, based on a review of the existing literature, the eastern wolf evolved in North America from a common ancestor with coyotes and now hybridizes with gray wolves where their ranges overlap. They cite phenotypic differences and concordant uniparentally inherited markers (Y chromosome and mitochondrial DNA (mtDNA)) as supporting a species-level distinction. Additionally, they note that geographic discontinuities in microsatellite data also indicate isolation and are consistent with spatially associated admixture. vonHoldt et al. (2011, 2016), using a high-density domestic dog SNP chip and whole-genome sequences, argue that the eastern wolf is instead a more recent lineage and is derived from the gray wolf. Their results indicate recent admixture with coyotes and show a geographic cline in the amount of coyote ancestry (Fig. 3). This cline can be explained by spatial patterns of wolf persecution by humans that result in lower population densities, decreasing the probability of finding a conspecific mate and thereby increasing the probability of admixture with other canid species. They argue that this recent admixture could be driving the phenotypic differences that are the primary evidence of the species designation by Chambers et al..
Determination of the historical context of admixture has implications for conservation. An admixed species is viewed differently if admixture is a natural part of its evolutionary history versus for a species where admixture is recent and driven by human activities. Formal model testing should be used to test specific hypotheses, as the data may be consistent with different hypotheses, each of which could have diverse implications for conservation and management. Additionally, haplotype analysis with whole-genome data will be necessary to estimate the timing of admixture events and new statistical approaches are needed to determine whether ancient hybridization, in addition to more recent admixture, was an important component of the evolutionary history of eastern wolves.
이 글은 위 출처의 원문을 그 라이선스(CC BY 4.0)에 따라 옮겼습니다. 참고문헌 목록과 인용 번호, 그림 · 표 · 각주 · 수식 · 부록은 빼고 절별로 나누어 실었으며, 그 밖의 문장은 원문 그대로입니다. 원 저작자와 발행 기관은 이 화면을 만들거나 검토하지 않았고, 이곳의 해석과 낱말 풀이는 공식 번역이 아닙니다.
