How genomics can help biodiversity conservation

Review How genomics can help biodiversity conservation Kathrin Theissinger, 1,36 Carlos Fernandes, 2,3,36 Giulio Formenti, 4,36 Iliana Bista, 5,6 Paul R. Berg, 7,8,9 Christoph Bleidorn, 10 Aureliano Bombarely, 11 Angelica Crottini, 12,13,14 Guido R. Gallo, 15 José A. Godoy, 16 Sissel Jentoft, 9 Joanna Malukiewicz, 17 Alice Mouton, 18 Rebekah A. Oomen, 8,9 Sadye Paez, 4 Per J. Palsbøll, 19,20 Christophe Pampoulie, 21 María J. Ruiz-López, 16,22 Simona Secomandi, 15 Hannes Svardal, 23 Constantina Theofanopoulou, 4,24 Jan de Vries, 25 Ann-Marie Waldvogel, 26 Guojie Zhang, 27,28,29 Erich D. Jarvis, 4 Miklós Bálint, 1 Claudio Ciofi, 30 Robert M. Waterhouse, 31,32 Camila J. Mazzoni, 33,34 Jacob Höglund, 35, *and The European Reference Genome Atlas Consortium 37 The availability of public genomic resources can greatly assist biodiversity assessment, conservation, and restoration efforts by providing evidence for scientifically informed management decisions. Here we survey the main approaches and applications in biodiversity and conservation genomics, considering practical factors, such as cost, time, prerequisite skills, and current shortcomings of applications. Most approaches perform best in combination with reference genomes from the target species or closely related species. We review case studies to illustrate how reference genomes can facilitate biodiversity research and conservation across the tree of life. We conclude that the time is ripe to view reference genomes as fundamental resources and to integrate their use as a best practice in conservation genomics. The value of integrating genomics into conservation We are in the midst of the sixth mass extinction, a biodiversity crisis with devastating consequences on ecosystem functioning and health, evolutionary heritage, and the adaptive potential of species, ultimately posing a major threat to humanity [1,2]. Although genetic diversity has long been recognized as fundamental to all levels of biological organization (individuals, populations, species, communities, and ecosystems), genomics is often neglected in biodiversity assessments and conservation efforts [3]. This gap between our ability to generate data and study genomic diversity and the use of genomics in conservation may be due to several factors [4]. These include prioritizing limited funds to address anthropogenic threats [5], limited knowledge transfer and collaboration between genomicists and conservation practitioners, stakeholders, and politicians [6], and a general lack of explicit references to measuring, monitoring, and preserving genetic diversity in the most relevant international regulations, such as those issued by the Convention for Biological Diversity [7] (although the situation may have recently started to change, https:// www.cbd.int/article/cop15-cbd-press-release-final-19dec2022). Biodiversity preservation critically depends on addressing key conservation issues. These include taxonomic identification and biodiversity monitoring associated with ecosystem protection and restoration (e.g., for invasive species management). At the same time, human activities exert significant demographic pressures on habitats and endangered species. This requires managing small populations, restoring and increasing genetic diversity of target species/populations, and supporting species adaptation to a changing environment. Genomic data can help tackle these issues as they allow us to characterize and monitor genetic diversity through a wide array of Highlights Genomics provides effective tools to characterize biodiversity, but the full implementation of genomic techniques in practical conservation is still limited. We review some of the main approaches in biodiversity genomics available to conservationists and genomicists. High-quality, long-read sequencing and bioinformatic technologies facilitate genome sequencing and assembly for any species. We summarize how reference genomes, in conjunction with population genomic data, can contribute to biodiversity monitoring, conservation, and restoration efforts. Over the past decade, many initiatives to generate reference genomes spanning the tree of life have emerged worldwide. We call for increased integration of reference genomes and population genomics data into interdisciplinary conservation efforts to fully unlock the potential of genomics in safeguarding global biodiversity. 1 LOEWE Centre for Translational Biodiversity Genomics, Senckenberg Biodiversity and Climate Research Centre, Georg-Voigt-Str. 14-16, 60325 Frankfurt/Main, Germany 2 CE3C - Centre for Ecology, Evolution and Environmental Changes & CHANGE - Global Change and Sustainability Institute, Departamento de Biologia Animal, Faculdade de Ciências, Universidade de Lisboa, 1749-016 Lisboa, Portugal Trends in Genetics, July 2023, Vol. 39, No. 7 https://doi.org/10.1016/j.tig.2023.01.005 545 © 2023 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Trends in Genetics OPEN ACCESS emerging tools. The novel insights that can be obtained from genomic data have led to the formation of several national and international initiatives aiming to expand the genomic resources available for non-model species (Table S1 in the supplemental information online). At the same time, scientists and practitioners are collaborating to: (i) standardize protocols for detecting and monitoring species genetic diversity and their adaptive potential, and (ii) integrate genetic and evolutionary knowledge into conservation planning [8]. These actions are critical to promoting transboundary management to ensure the persistence of populations and species and, ultimately, the continued provision of nature-based ecosystem services. Recently, in a forum article [9]wearguedforanew era in conservation genomics underpinned by reference genomes (see Glossary), also taking advantage of the increasing ease and widespread interest in generating them. Here, we expand this perspective in more detail, illustrated with recent examples. We also provide an overview on how genomics can help biodiversity conservation, which should be helpful to both conservationists who are not experts in genomics and to genomicists without a conservation biology background interested in engaging with conservation genomics research. Genomic approaches in biodiversity research High-throughput genomic sequencing technologies have lately evolved from mainly generating short (50–300 bp) to much longer (>10 000 bp) DNA sequencing reads. Genetic and genomic approaches commonly used in biodiversity research include DNA barcoding/metabarcoding, reduced representation DNA techniques, transcriptome sequencing (RNA-Seq), and wholegenome (re-)sequencing (Figure 1). Successful execution of these approaches depends on the quantity and quality of the available biological material, laboratory and bioinformatic skills, feasibility and costs, and on the quality and completeness of the available reference genome database. We review these common approaches, including their specific merits and limitations (Table 1). Since the quality of the starting biological material is a major consideration in applying genomic approaches, we also review approaches that use noninvasive sampling. DNA barcoding and metabarcoding DNA barcoding has become a standard, efficient genetic approach for species identification and biodiversity monitoring [10]. DNA barcoding sequences informative DNA loci with universal or taxon-specific primers that anneal to conserved flanking regions. Initially, DNA barcoding was based on the mitochondrial cytochrome coxidase subunit I (COI) gene and mainly focused on animals, particularly invertebrates. Over the years, additional DNA loci have been utilized to barcode vertebrates (e.g., 12S, 16S, and Cytb in mtDNA), plants (rbcL and matK in cpDNA), fungi (ITS in rDNA), protists and nematodes (18S in rDNA), and bacteria (16S in rDNA) [10]. DNA metabarcoding combines the principles of DNA barcoding with next-generation sequencing (NGS), enabling the analysis of complex samples containing a mixture of specimens and/or species [11]. Metabarcoding has been widely used in biodiversity assessment and monitoring (e.g., species turnover during ecosystem restoration [12], mapping of ecological networks [13], or detection of invasive species [14]). A key advantage of metabarcoding is bulk sampling and sequencing, circumventing costly sorting and processing of samples into individual specimens (Table 1), thereby enabling high-throughput ecosystem-wide assessments and monitoring in most environments [15]. Moreover, metabarcoding is an appropriate approach to sequence environmental DNA, often degraded into short fragments in an environmental medium [16]. However, the short length of the DNA regions targeted in barcoding and metabarcoding can often limit accurate characterization of the genetic and taxonomic diversity in a community, failing to discern closely related taxa or taxa with introgressed nuclear genes or organellar genomes. Genome skimming circumvents some of the experimental biases in metabarcoding, potentially allowing more accurate metagenomic estimates of biodiversity [17,18] and wildlife forensic investigations [19]. Genome skimming has been conducted in situ, enabling rapid, field-based 3 Faculdade de Psicologia, Universidade de Lisboa, Alameda da Universidade, 1649-013 Lisboa, Portugal 4 The Rockefeller University, 1230 York Ave, New York, NY 10065, USA 5 Naturalis Biodiversity Center, Darwinweg 2, 2333, CR, Leiden, The Netherlands 6 Wellcome Sanger Institute, Tree of Life, Wellcome Genome Campus, Hinxton, CB10 1SA, UK 7 NIVA - Norwegian Institute for Water Research, Økernveien, 94, 0579 Oslo, Norway 8 Centre for Coastal Research, University of Agder, Gimlemoen 25j, 4630 Kristiansand, Norway 9 Centre for Ecological and Evolutionary Synthesis, Department of Biosciences, University of Oslo, PO BOX 1066 Blinderm, 0316 Oslo, Norway 10 University of Göttingen, Department of Animal Evolution and Biodiversity, Untere Karspüle, 2, 37073, Göttingen, Germany 11 Universita degli Studi di Milano, Via Celoria 26, 20133, Milan, Italy 12 CIBIO/InBio, Centro de Investigação em Biodiversidade e Recursos Genéticos, Rua Padre Armando Quintas, 7, 4485-661, Portugal 13 Departamento de Biologia, Faculdade de Ciências, Universidade do Porto, 4099-002 Porto, Portugal 14 BIOPOLIS Program in Genomics, Biodiversity and Land Planning, CIBIO, Campus de Vairão, 4485-661 Vairão, Portugal 15 Department of Biosciences, University of Milan, Milan, Italy 16 Estación Biológica de Doñana, CSIC, Calle Americo Vespucio 26, 41092, Sevillle, Spain 17 Primate Genetics Laborator, German Primate Center, Kellnerweg 4, 37077, Göttingen, Germany 18 InBios - Conservation Genetics Lab, University of Liege, Chemin de la Vallée 4, 4000, Liege, Belgium 19 Groningen Institute of Evolutionary Life Sciences, University of Groningen, Nijenborgh, 9747, AG, Groningen, The Netherlands 20 Center for Coastal Studies, 5 Holway Avenue, Provincetown, MA 02657, USA 21 Marine and Freshwater Research Institute, Fornubúðir, 5,220, Hanafjörður, Iceland 22 CIBER de Epidemiología y Salud Pública (CIBERESP), Spain 23 Department of Biology, University of Antwerp, Universiteitsplein 1, 2610 Wilrijk, Antwerp, Belgium 24 Hunter College, City University of New York, NY, USA Trends in Genetics OPEN ACCESS 546 Trends in Genetics, July 2023, Vol. 39, No. 7 assessments of non-model organisms (e.g., [20]) (Table 1). Use of genome skimming can also provide increased information content compared with barcoding and can be applied for sequencing of challenging specimens, such as museum specimens whose DNA may be too degraded and challenging for Sanger sequencing [21]. Reduced genomic representation Reduced representation DNA sequencing (RRS) approaches are predominant in studies aimed at non-model species [22]. RRS approaches provide genome-wide data in large sample sizes at a comparably reduced cost (Table 1). A small fraction of the genome is reproducibly targeted in each specimen, either using restriction endonucleases and size selection (e.g., RADseq and related methods [23]) or captured by hybridization using probes (e.g., ultra-conserved elements or custom baits) or a combination of both (e.g., HyRAD [24], Rapture [25]). Although RRS approaches only capture a small portion of the genome, they provide sufficient genome-wide data to estimate genetic diversity, inbreeding,effective population size, population structure and assignment, gene flow, phylogeographic patterns, and phylogenetic relationships [22,26]. However, combining RRS-derived data from different studies necessitates identical experimental protocols, which limits replicability. Moreover, RRS approaches utilizing restriction endonucleases (and PCR) may be subject to allelic dropout [23]. Although RRS approaches can be conducted without a reference genome, alignment to a reference genome improves inferences obtained from RRS data [27]. Furthermore, a reference genome provides genome coordinates for most SNPs, thereby facilitating the identification of linked loci, which is key to many population genetic inferences [28]. Gene expression Gene expression data (usually RNA-Seq) have given rise to a new conservation framework by characterizing genetic variationinnaturalpopulationsthrough functional variation [29] and rapid responses of individuals or populations to environmental change [30]. Differences in gene expression have been linked to life history traits and population dynamics, aiding in the identification of candidate genes potentially affecting eco-evolutionary processes [31]. Gene expression data have provided insights into responses to pesticide exposure [32] and susceptibility or resistance to diseases [33]. Gene expression data have also been used to predict range shifts and identify vulnerable populations or adaptive phenotypes [34]. Further, gene expression data have been used to identify entire gene networks in the absence of prior knowledge of the genes involved [35](Table 1). Analyzing gene expression data is more challenging than DNA-based analyses, because RNA is more susceptible to degradation than DNA and because transcription varies across cell and tissue types, sex, age, physiological and life stage, and even according to circadian rhythms. Consequently, gene expression studies require careful consideration of confounding biological factors [36]. As for RRS approaches, gene expression data are most informative when aligned against a reference genome [37], which provides an independent functional annotation to support the assignment of reads to genes [38,39]. In parallel with genomic assessments of DNA and RNA, epigenomics is being increasingly applied in biodiversity conservation research [40]. Epigenetic modifications may play a key role in phenotypic and biological innovations during major ecological transitions, thus potentially enabling rapid evolutionary responses [41,42]. Identifying the epigenetic modifications that modulate phenotypic variation can therefore potentially identify species that are vulnerable to environmental changes and be informative for monitoring the evolution of invasive species [43]. These approaches rely entirely on the availability of a reference genome to identify epigenetic signatures. Whole-genome sequencing Whole-genome sequencing (WGS) data offer unparalleled power and resolution in analyses of demographic history, admixture and introgression, recombination and linkage disequilibrium, 25 University of Goettingen, Institute for Microbiology and Genetics, Department of Applied Bioinformatics, Goettingen Center for Molecular Biosciences (GZMB), Campus Institute Data Science (CIDAS), Goldschmidtstr. 1, 37077, Goettingen, Germany 26 Institute of Zoology, University of Cologne, Zülpicherstrasse 47b, D-50674, Cologne, Germany 27 Evolutionary & Organismal Biology Research Center, Zhejiang University School of Medicine, Hangzhou, 310058, China 28 Villum Center for Biodiversity Genomics, Section for Ecology and Evolution, Department of Biology, University of Copenhagen, Denmark 29 State Key Laboratory of Genetic Resources and Evolution, Kunming Institute of Zoology, Chinese Academy of Sciences, Kunming, 650223, China 30 University of Florence, Department of Biology, Via Madonna del Piano 6, Sesto Fiorentino, (FI) 50019, Italy 31 University of Lausanne, Department of Ecology and Evolution, Le Biophore, UNIL-Sorge, 1015 Lausanne, Switzerland 32 Swiss Institute of Bioinformatics, 1015 Lausanne, Switzerland 33 Leibniz Institute for Zoo and Wildlife Research (IZW), Alfred-Kowalke-Str 17, 10315 Berlin, Germany 34 Berlin Center for Genomics in Biodiversity Research (BeGenDiv), Koenigin-Luise-Str 6-8, 14195 Berlin, Germany 35 Department of Ecology and Genetics, Uppsala University, Norbyvägen 18D, 75246, Uppsala, Sweden 36 The authors are co-first authors. 37 The ERGA Consortium. *Correspondence: [email protected] (J. Höglund). Trends in Genetics OPEN ACCESS Trends in Genetics, July 2023, Vol. 39, No. 7 547 genetic load, natural selection, and species diversification [44–47]. Elucidating evolutionary processes may require analyses of linkage disequilibrium between many physically distant loci [48]. WGS data enable detection of genomic regions under selection, mutations in regulatory elements, rare variants and structural variation, and investigating the genetic basis and architecture of phenotypic traits (e.g., disease susceptibility/resistance in endangered species) [49]. WGS data can also provide unique insights into phylogenetic relationships, evolutionary history, extinct and cryptic taxa, and ancestral intraspecific genetic diversity and structure, which are valuable in assessing temporal genomic erosion and guiding conservation and restoration efforts [50,51]. Since samples of extinct or endangered species usually are rare, WGS data maximize the genomic information obtained from each sample. The highly fragmented DNA in museum and subfossil specimens (often <100 bp) requires the availability of reference genomes for WGS read mapping and downstream evolutionary analyses [52]. A cost-efficient approach for population genomic studies is Pool-Seq (i.e., deep sequencing of multiple pooled individuals) [53](Table 1). Pool-Seq has been applied to investigate genomewide patterns of genetic diversity, signatures of adaptation, population differentiation, and genotype–environment associations [54–56]. In some cases, Pool-Seq is the only possible or practical approach, for example, with microscopic organisms. It facilitates surveys of multiple populations and conditions, with large numbers of individuals (>50) sampled in each, and can significantly reduce costs in species with a small to intermediate-sized genome (<1 Gb). Despite the cost-saving advantages, Pool-Seq entails some analytical and statistical challenges [53]. The robustness and accuracy of results directly depend on the pool size [53] and can be reduced in complex systems with patterns of weak population structure [57]. Noninvasive genomic sampling Noninvasive or minimally invasive sampling of biological material (e.g., from feces, feather, or hair) [58,59] is commonly used in wildlife monitoring as the main source of genetic material to provide insights into the ecology of endangered species [60]. Noninvasive samples typically contain low amounts of poor-quality DNA, often contaminated with exogenous DNA [59]. The technical challenges specific to noninvasive samples have been addressed through approaches that maximize the amount of retrieved endogenous DNA, such as: (i) SNP genotyping arrays [61]; (ii) multiplex PCR amplicon sequencing [62]; or (iii) target sequence capture using RNA/DNA oligonucleotides [63], which may be developed from existing information obtained using RRS approaches or reference genomes generated from high-quality DNA specimens. What can reference genomes bring to biodiversity conservation? Since reference genomes are a prerequisite or enhance many of the analyses outlined earlier, their availability is desirable for any biodiversity genomics project [64]. Indeed, reference genomes are becoming commonplace across the entire tree of life. This is due to recent technological developments (Box 1) that enable the generation of accurate, chromosome-level genome assemblies for non-model organisms [9], and to multiple concerted international initiatives devoted to systematically generating reference genomes at scale (Table S1 in the supplemental information online). Next, we summarize how the availability of reference genomes substantially improves the use of genomics in biodiversity conservation. The key to the whole trove of genomic information Reference genomes provide the backbone and framework to map and annotate intraand interspecific genomic variation [65–67]. They also provide a context to genetic variants, regions of low heterozygosity, patterns of linkage disequilibrium, and site frequency spectra (SFS) along chromosomes. In turn, these can be used to identify candidate genes associated with Trends in Genetics OPEN ACCESS 548 Trends in Genetics, July 2023, Vol. 39, No. 7 Glossary Accessory genome: portion of the pangenome of a species containing the genes that are not present in all lineages. Admixture: production of new genetic combinations in hybrid populations through recombination. Allelic dropout: experimental failure to detect allele(s) at a heterozygous locus. Chromosome-level assembly: the process of generating a contiguous sequence of all chromosomes of a genome, often aided by genetic maps or proximity ligation techniques (3C-seq, Hi-C); term also used to refer to the resulting genome sequence. Core genome: portion of the pangenome of a species containing the genes present in all lineages. De novo assembly: the process of generating a genome sequence from individual sequencing reads without the use of an existing reference; the term is also used to refer to the resulting genome sequence. Ecotypes: intraspecific phenotypes adapted to different ecological niches. Effective population size: N e ;key parameter that reflects the rate of genetic drift and inbreeding; the size of an idealized population with the observed rate of genetic drift. Genetic load: actual or potential reduction in mean population fitness due to genetic causes (deleterious mutations, genetic drift, inbreeding, migration, recombination). Genome skimming: low-depth wholegenome sequencing typically used to recover high-copy DNA regions, such as mitochondrial and chloroplast DNA and repetitive nuclear loci (e.g., satellite DNA, transposable elements). Genomic erosion: loss and degradation of genome-wide diversity due to inbreeding, genetic drift, introgression, and selection in small or isolated populations. Hybridization: interbreeding of individuals from genetically distinct populations. Inbreeding: production of offspring from the mating of closely related individuals. Inferred more formally from the inbreeding coefficient (F), which denotes the probability that two gene copies at a locus are identical by descent. Inbreeding depression: reduced fitness in offspring as a result of inbreeding. phenotypes and adaptations and elucidate complex genomic features such as satellite DNA, gene families, segmental duplications, or structural variation. Structural variation, such as large inversions that can limit recombination, may be linked to adaptation to environmental change and may have important monitoring and management implications (Box 2: Case study 1 on the Atlantic cod) [68]. Chromosomal-level genome assemblies improve detection of structural variants of different sizes, including some spanning megabases across multiple gene families (Box 2: Case study 2 on the horseshoe crabs). Reference genomes can reveal integration, recombination, or deletion of mobile elements, which may play a role in genomic and taxonomic diversification [69]. Reference genomes also enable the rapid and efficient in silico development of custom SNP or STR assays targeting specific functional variants of interest, which facilitates the analysis of noninvasive, forensic or museum DNA samples [70]. Functional and adaptive genetic variation Reference genomes facilitate the identification of functional genetic variation underlying phenotypic traits, fitness variation, and adaptive potential [71–73]. Such analyses are mostly successful at identifying variants with large effects [71,74,75], likely representing a small proportion of the adaptive functional variation. However, large-effect variants can underlie traits that are responsible for resilience to anthropogenic threats such as diseases or climate change. Such traits, identified, for example, in genome-wide association studies (GWAS), can be selected for in captive breeding or translocation programs (Box 2: Case study 3 on the ash dieback). Those programs must then carefully consider issues such as maintaining genome-wide diversity and evolutionary potential [76–78] and genetic adaptation to captivity [79]. Most phenotypic traits are polygenic (i.e., under the control of multiple interacting loci with small effects). These loci are, in turn, likely subject to gene regulatory networks difficult to fathom with DNA-based tests aimed at detecting adaptive variation at individual loci. To this end, statistical tools have been developed to detect contributions from multiple loci across the entire genome that may underlie the function and heritability of trait variation [80], as well as more direct approaches such as massively parallel reporter assays [81]. Reference genomes facilitate the estimation of the relative fitness of alleles and genotypes in threatened species [82], which may not be amenable to experimental approaches, for example, by WGS comparisons of individuals with different phenotypes [83]. Notwithstanding the analytical challenges, WGS data, combined with phenotypic and fitness data, enable a thorough understanding of adaptive variants [74] that drive responses (or the lack thereof) to current and future anthropogenic threats [84], such as ongoing environmental change [85]. Incorporating insights of adaptive potential into species distribution models helps modeling climate change effects on endangered populations [86,87], such as genotype–phenotype associations underlying species resilience, which can guide assisted breeding programs [88]( Box 2:C ase study 4 on the European beech). Identifying adaptive variation may also aid in defining conservation units that ensure the preservation of evolutionary heritage and adaptive potential [89]. Inbreeding and genetic load Assessing detrimental fitness effects associated with inbreeding (inbreeding depression)isa central topic in conservation genetics [90,91]. The availability of annotated reference genomes is invaluable for characterizing inbreeding and the genetic architecture of inbreeding depression, in particular the number of loci involved, their effect size, and the contributions of deleterious recessive alleles [92,93]. This knowledge can guide the selection of founders, breeders, and individuals for reintroduction in ex situ conservation management and help design purging strategies, or inform monitoring of deleterious variants and genomic erosion in the wild [94]. Runs of homozygosity (ROH) analyses are useful for distinguishing historical andcontemporary inbreeding, pinpointing candidate loci contributing to inbreeding depression (particularly loci with large effects), and selecting captive breeders with different ROH to minimize inbreeding Trends in Genetics OPEN ACCESS Trends in Genetics, July 2023, Vol. 39, No. 7 549 Introgression: gene flow between hybridizing populations or species by backcrossing of hybrids with one or both parental species. Linkage disequilibrium: nonrandom association between alleles at different loci. Mobile elements: DNA sequences that are able to integrate into new sites within a genome, sometimes also transferred among species. They are one of the major drivers of genome evolution. Outbreeding: mating between individuals from genetically distinct lineages (e.g., populations, subspecies, or species). Outbreeding depression: reduced fitness of offspring from matings between genetically divergent individuals. Pangenome: the entire set of DNA sequences (or genes) of a species represented by the core genome and the accessory genome. Reference genome: contiguous and accurate genome assembly representative of a species with coordinates of genes and other important features annotated. Restriction endonuclease: an enzyme produced by certain bacteria that has the property of cleaving DNA molecules at or near a specificsequence motif. Runs of homozygosity (ROH): genome regions that are homozygous due to identity-by-descent. ROH arise when two copies of an ancestral haplotype are brought together in the same individual. The number and size of ROHs are the result of various processes, in particular nonrandom mating, inbreeding, and population demographic history. Site frequency spectrum (SFS): distribution of allele frequencies of a given set of loci (often SNPs) in a population sample (aka ‘allele frequency spectrum’). Structural variation: regions of a chromosome presenting structural changes such as insertion, deletion, inversion, or translocation of DNA. Transposable elements: see ‘Mobile elements’. depression [90,91,95]. Reference genomes facilitate predicting the putative functional significance and fitness consequences of specific allelic variants that have drifted to high frequency or fixation in small wild populations and estimating the accumulated genetic load and dynamics of deleterious variation [51,96–98]. Such in-depth reference genome-based analyses are being carried out in an increasing number of endangered species (e.g., Box 2: Case study 5 on the Iberian lynx, the crested ibis [99], and the Sumatran rhinoceros [92]). In perspective, reference genomes combined with WGS data provided invaluable insights into the long-term management of translocated populations (Box 2: Case study 6 on the Florida panther). Outbreeding and hybridization Outbreeding can also result in a reduction of fitness, known as outbreeding depression, due to chromosomal or genic incompatibilities, epistatic interactions, disruption of interactions between coadapted genes, or introduction of variants that are maladaptive to local environmental conditions. In risk assessments of outbreeding depression, WGS and reference genomes can be invaluable in enabling detailed analyses of adaptive divergence and structural variation across thegenome[46,82]. This knowledge is important for reducing the chances of outbreeding depression when selecting source populations and individuals for translocations to reinforce a population threatened with extinction [100]. Human-mediated hybridization and introgression have increased dramatically worldwide. In association with the perturbation and homogenization of environments, hybridization can become a major conservation concern, threatening biodiversity and evolutionary heritage. Genome-wide data make it possible to characterize introgression patterns, dynamics, and admixture proportions more effectively than traditionally possible with Transcriptome Reduced representation Barcode or or Whole genome (chromosome-scale /complete) Whole genome (fragmented/draft) Native genome Sequencing & assembly Chromosome 1 Chromosome 2 Chromosome n Organellar DNA Exon1 Exon2 Exon3 Exon1 Exon2 Exon4 Practicability for broad biodiversity Exon3 Exon1 Exon2 Exon3 Exon1 Exon2 Exon4Exon3 Information content Trends Trends in in Genetics Genetics Figure 1. Genomic approaches for biodiversity research. In a eukaryotic cell, DNA is natively packed and allocated over different chromosomes, ranging from a few to, more rarely, hundreds in the nucleus, as well as over organellar genomes (mitochondrial DNA and, in plants and algae, also plastid DNA), making up the genome. Current technology cannot sequence entire chromosomes longer than the read lengths (usually 10–100 kbp). Such reads yield short pieces of sequence information. The shortest pieces used for biodiversity research are the amplification and/or selective sequencing of specific regions of the genomes: multiple loci (‘multilocus’)or single marker regions (‘barcodes’). Reduced representation samples low copy regions of the DNA. Transcriptomic data, in particular RNA-Seq at a genome-wide scale, can be assembled de novo or mapped onto a genomic backbone. Finally, reads are assembled into highly contiguous chromosomes (chromosome scale/ complete) or less contiguous (fragmented/draft) pseudomolecules. On the right, arrows give an indication of (i) the ease of covering a broad taxonomic diversity with a given technique, and (ii) the amount of biodiversity or population information produced by a technique. Trends in Genetics OPEN ACCESS 550 Trends in Genetics, July 2023, Vol. 39, No. 7 Table 1. List of genomic approaches with application and comparison of raw sequencing costs (i.e., costs of sample collection, researcher time, and analysis are not included) DNA barcoding/ metabarcoding Genome skimming Reduced representation DNA sequencing Transcriptome sequencing Whole-genome resequencing What genome do you get? None Organelle, k-mer representation of nuclear None Coding regions only, variable fragmentation Nonrepetitive genome, depends on coverage Cost in dollars (as of date) a $5 per sample (Sanger/NGS) $50 $50 $100–$400 $100–$600 What type of samples are needed Fresh tissue samples, museum specimens, noninvasive samples Fresh tissue samples, museum specimens Fresh tissue samples, museum specimens, noninvasive samples Tissue-specific, live/fresh, flash frozen/in RNA buffer Fresh tissue samples, museum specimens Genetic diversity b Yes, but limited Yes, but limited Yes Yes Yes Population structure c Yes, but weak to detect shallow/cryptic genetic structure; economical for detailed spatial sampling Yes, typically organelle based Yes Yes Yes Phylogenetic information Yes, but barcode based Yes, typically organelle based Yes Yes Yes Introgression event No No Yes, but no individual genes Yes, but limited detection power Yes QTL mapping No No Yes, but low resolution Yes, expression QTL (eQTL) Yes Natural selection signal detection No Yes, on organelle genes Yes Yes Yes Gene structure study No Yes, on organelle genes Potentially, if reference genome available No Yes, if reference genome available Gene family analyses No Yes, on organelle genes Potentially, if reference genome available Yes Yes, if reference genome available Genome rearrangement study No Yes, on organelle Potentially, if reference genome available No Yes, if reference genome available Functional genomic study d No Yes, on organelle genes No Yes No Genome size estimation No Yes, typically organelle genome No No Depending on coverage Linkage disequilibrium No No Yes, usually, not always Limited Yes Demographic reconstructions (MSMC) e No No No No Yes Demographic reconstructions from SFS No No Yes No Yes GWAS No No Yes, but low resolution Yes, T(ranscriptome) WAS Yes Whole-genome sequencing Short reads Linked reads (10X/stLFR/Tell-Seq) Long reads + scaffolding technologies What genome do you get? Very fragmented, contig N50 <10 kb Contig N50 >200 kb, scaffold N50 >10 Mb Reference genome Cost in dollars (as of date) a $100–$600 $2000 $5000 What type of samples are needed Fresh tissue samples, museum specimens Frozen upon collection Large quantity of high molecular weight DNA (continued on next page) Trends in Genetics OPEN ACCESS Trends in Genetics, July 2023, Vol. 39, No. 7 551 only dozens of markers [101]. Introgression rates in different genomic regions can vary. Wholegenome data mapped to reference genomes are useful to identify admixture (or migrant) tracts along individual genomes and thus to understand local effects of introgression and the age of admixture events [102,103]. Genome editing: engineering adaptation, gene drives, and de-extinction Genome-editing tools, such as CRISPR-Cas9 [104], allow precise modification of genes and genomes of living organisms. Genome editing can be applied as a tool to mediate locusspecific genetic rescue in endangered species threatened with high frequencies of deleterious Table 1. (continued) Whole-genome sequencing Short reads Linked reads (10X/stLFR/Tell-Seq) Long reads + scaffolding technologies Genetic diversity b Yes Yes Yes Population structure c Yes Yes Yes Phylogenetic information Yes Yes Yes Introgression event Yes, but limited detection power Yes Yes QTL mapping Yes, but low resolution Yes Yes Natural selection signal detection Partially, if reference genome available Yes Yes Gene structure study Partially, if reference genome available Yes, miss few regions Yes Gene family analyses No Yes Yes Genome rearrangement study No Yes, but fragmented karyotype Yes, can trace karyotype evolution Functional genomic study d Yes, but missing many regions Yes, but missing some regions Yes Linkage disequilibrium Yes Yes Yes Demographic reconstructions (MSMC) Yes Yes Yes Demographic reconstructions from SFS Yes Yes Yes GWAS Yes Yes Yes a Based on a 2 Gb vertebrate genome. b Noting that single genomes deliver estimates from two haplotypes only. c Noting that single specimens do not deliver to this aim well. d Integrated with ChIP-seq, Hi-C, RNAseq for regulatory element annotation. e Multiple sequentially Markovian coalescent. Box 1. Genomes, population genomics, and pangenomes Reference genomes, point representations of the structure and organization of a species’genome, are expected to revolutionize conservation genomics [9]. Currently, a combination of single-molecule long-read sequencing (SMRT or nanopore sequencing) or linked reads (e.g., TELL-seq or stLFR) for contigging, optical maps, and/or proximity ligation reads (e.g., 3C-seq or Hi-C) for scaffolding appears to be a general strategy capable of generating phased chromosomescale reference genomes across the tree of life [124,125](seeTable 1 in main text). When sufficient high molecular weight DNA for long-read sequencing remains problematic to obtain (e.g., with small organisms, albeit solutions are rapidly becoming available [126]), an alternative cost-effective strategy is represented by using linked reads only [127]. To gain more accurate insights into a species’genomic diversity, the genomes of large numbers of individuals can be resequenced. Such multiple genomes do not have to be assembled de novo (de novo assembly) and may be derived from cheaper and quicker short-read sequencing and subsequent mapping to an existing assembly [128]. Additionally, the availability of multiple reference genomes per species provides a more complete genome characterization [129,130] and allows estimating a species pangenome [131]. Individuals of a species are part of a pangenomic architecture, sharing a larger proportion of their genes as a core set, whilst differing in a smaller proportion of dispensable genes [132]. The analysis of such data helps to differentiate adaptive and neutral evolutionary processes, to define populations, ecotypes, and species, and identify core and accessory genomic regions, gene flow, hybridization, and incomplete lineage sorting [133]. Trends in Genetics OPEN ACCESS 552 Trends in Genetics, July 2023, Vol. 39, No. 7 Box 2. Case studies 1. Atlantic cod The Atlantic cod (Gadus morhua;Figure IA) comprises several cryptic ecotypes differing in feeding, spawning, and migratory behaviors and, consequently, in population dynamics relevant for fisheries management [134]. The Atlantic cod was one of the first non-model organisms with a chromosome-anchored genome assembly [135]. Combined with population genomic analyses, the genome assembly enabled identification of four large chromosomal inversions that discriminate between the migratory and nonmigratory ecotypes in Norwegian and Icelandic waters and are associated with adaptation to environmental conditions [134,136]. Each of the four inversion regions contains hundreds of genes with a high level of divergence between the inverted and noninverted haplotypes. The four inversions originated independently up to 1.7 million years ago and have been maintained due to strong selection [136,137]. 2. Horseshoe crabs Horseshoe crabs (Figure IB) belong to the chelicerate order Xiphosura. This enigmatic order comprises only four extant species (Limulus polyphemus,Tachypleus gigas, Tachypleus tridentatus, and Carcinoscorpius rotundicauda), all considered ‘living fossils’since their morphology is virtually identical to that of their Triassic ancestors. Recent studies have generated chromosome-level assemblies of three of these species, and analyses suggest three rounds of whole-genome duplication [138,139]. Several gene families, particularly those involved in innate immunity, have undergone extensive tandem duplication. These expanded gene families may be important components of the innate immune system of horseshoe crabs, whose amebocytes are presently exploited for detecting endotoxin contamination [139]. These genomic resources are of value for breeding programs and conservation [138]. 3. Ash dieback In the mid-1990s, European ash trees (Fraxinus excelsior;Figure IC) started dying across Europe from a new disease, ash dieback, which spread rapidly and was determined to be caused by Hymenoscyphus fraxineus, a fungal pathogen new to Europe. Genome sequencing of this fungus along with isolates from East Asia confirmed that the European outbreak was likely caused by the introduction of imported Asian Fraxinus mandshurica. Importantly, the fungal genomes showed that at most two independent introductions of H. fraxineus had occurred [140]. The availability of the fungal genomes will aid surveillance for new introductions. While no ash trees appeared to be fully resistant, threatening extinction of this iconic species, analysis of Danish forestry germplasm identified individuals that had reduced susceptibility. An international consortium sequenced the European ash genome [141] and used it to map polygenic traits associated with reduced susceptibility [142], promising to assist in directed breeding to restock ash forests across the continent. 4. European beech In the summers of 2018 and 2019, around two-thirds of European beech trees (Fagus sylvatica;Figure ID) were damaged or killed by extreme drought. It is critical to keep beech woods healthy, as they are home to over 6000 other species of animals and plants. However, not all the trees in each forest responded in the same way, with severely damaged trees often sitting less than 5 m from fully healthy ones. This suggests that the genetic makeup of a tree, rather than its local environment, determines how well it copes with drought. After the beech genome was assembled [143], over 400 European beech trees from pairs of neighboring trees that had responded differently to the droughts were analyzed [88]. More than 80 regions of the genome differed between healthy and damaged trees. Subsequently, a genetic test was developed which can quickly and inexpensively predict how well an individual beech tree might cope with drought. Now this genetic test can be used to select and reproduce trees that are better adapted to drought. 5. Iberian lynx One of the first whole genomes sequenced for a highly endangered species was that of the Iberian lynx (Lynx pardinus;Figure IE). Whole-genome resequencing analyses revealed that the species underwent a history of serial population declines, leading to one of the lowest genomic diversities ever reported [65]. Historical genetic data allowed the quantification of recent diversity loss and validated the mixing of the two differentiated remnant populations [144]. Subsequent genome-wide studies focused on functional variation have shown a higher genetic load in the more bottlenecked lynx populations [145], but also the purging of highly deleterious variants in Iberian lynx with respect to the more abundant Eurasian lynx [146]. These studies have also generated a catalog of potentially deleterious variation that can now be tested for association with reduced fitness and genetic diseases. The access to genome-wide variation also allowed the selection of a set of highly efficient and informative SNPs [147], which is now being extensively applied in combination with noninvasive samples to monitor and manage genetic diversity across ex situ, remnant, and reintroduced populations. These concerted conservation efforts aided by genetic information succeeded in avoiding the imminent extinction and in reverting the negative population trends of the species by the turn of the century. 6. Florida panther In the early 1990s, isolated Florida panthers, a population of puma (Puma concolor;Figure IF) in the Big Cypress National Preserve, showed low reproductive success and multiple signs of inbreeding. Therefore, pumas from Texas were translocated to support the isolated Florida population. Genetic diversity and reproductive success increased and the consequences of severe inbreeding, such as undescended testicles, heart failures, cowlick, and kinked tails, vanished [148]. A recent study explored the genomic consequences of the long-term isolation of this and other puma populations using a reference genome and a geographically broad panel of resequenced individuals [95]. The genome of a Florida panther descended from translocated Central American individuals had long tracts of homozygosity despite recent outbreeding. This suggests that sustaining diversity in small and isolated populations will require either repeated translocations or restoration of landscape connectivity. 7. Freshwater sponge Relatively few chromosome-level reference genomes are available for non-model invertebrate species, which hampers the implementation of conservation or biodiversity genomics approaches in a group of species that spans more than 95% of the animal tree of life. The freshwater sponge Ephydatia muelleri (Figure IG) is a common species in Palearctic lakes and freshwater streams that can even colonize drinking water systems, is transported by migratory birds and plays important ecological roles where it Trends in Genetics OPEN ACCESS Trends in Genetics, July 2023, Vol. 39, No. 7 553