Experimental support for genomic prediction of climate maladaptation using the machine learning approach Gradient Forests

Local adaptation of Pinus leiophylla under climate and land use change models in the Avocado Belt of Michoacán

Climate change and land use change are two main drivers of global biodiversity decline, decreasing the genetic diversity that populations harbour and altering patterns of local adaptation. Landscape genomics allows measuring the effect of these anthropogenic disturbances on the adaptation of populations. However, both factors have rarely been considered simultaneously. Based on a set of 3660 SNPs from which 130 were identified as outliers by a genome-environment association analysis (LFMM), we modelled the spatial turnover of allele frequencies in 19 localities of Pinus leiophylla across the Avocado Belt in Michoacán state, Mexico. Then, we evaluated the effect of climate change and land use change scenarios, in addition to evaluating assisted gene flow strategies and connectivity metrics across the landscape to identify priority conservation areas for the species. We found that localities in the centre-east of the Avocado Belt would be more vulnerable to climate change, while localities in the western area are more threatened by land conversion to avocado orchards. Assisted gene flow actions could aid in mitigating both threats. Connectivity patterns among forest patches will also be modified by future habitat loss, with central and eastern parts of the Avocado Belt maintaining the highest connectivity. These results suggest that areas with the highest priority for conservation are in the eastern part of the Avocado Belt, including the Monarch Butterfly Biosphere Reserve. This work is useful as a framework that incorporates distinct layers of information to provide a more robust representation of the response of tree populations to anthropogenic disturbances.

Residual networks without pooling layers improve the accuracy of genomic predictions

KEY MESSAGE

Residual neural network genomic selection is the first GS algorithm to reach 35 layers, and its prediction accuracy surpasses previous algorithms. With the decrease in DNA sequencing costs and the development of deep learning, phenotype prediction accuracy by genomic selection (GS) continues to improve. Residual networks, a widely validated deep learning technique, are introduced to deep learning for GS. Since each locus has a different weighted impact on the phenotype, strided convolutions are more suitable for GS problems than pooling layers. Through the above technological innovations, we propose a GS deep learning algorithm, residual neural network for genomic selection (ResGS). ResGS is the first neural network to reach 35 layers in GS. In 15 cases from four public data, the prediction accuracy of ResGS is higher than that of ridge-regression best linear unbiased prediction, support vector regression, random forest, gradient boosting regressor, and deep neural network genomic prediction in most cases. ResGS performs well in dealing with gene-environment interaction. Phenotypes from other environments are imported into ResGS along with genetic data. The prediction results are much better than just providing genetic data as input, which demonstrates the effectiveness of GS multi-modal learning. Standard deviation is recommended as an auxiliary GS evaluation metric, which could improve the distribution of predicted results. Deep learning for GS, such as ResGS, is becoming more accurate in phenotype prediction.

Incorporating adaptive genomic variation into predictive models for invasion risk assessment

Global climate change is expected to accelerate biological invasions, necessitating accurate risk forecasting and management strategies. However, current invasion risk assessments often overlook adaptive genomic variation, which plays a significant role in the persistence and expansion of invasive populations. Here we used Molgula manhattensis, a highly invasive ascidian, as a model to assess its invasion risks along Chinese coasts under climate change. Through population genomics analyses, we identified two genetic clusters, the north and south clusters, based on geographic distributions. To predict invasion risks, we employed the gradient forest and species distribution models to calculate genomic offset and species habitat suitability, respectively. These approaches yielded distinct predictions: the gradient forest model suggested a greater genomic offset to future climatic conditions for the north cluster (i.e., lower invasion risks), while the species distribution model indicated higher future habitat suitability for the same cluster (i.e, higher invasion risks). By integrating these models, we found that the south cluster exhibited minor genome-niche disruptions in the future, indicating higher invasion risks. Our study highlights the complementary roles of genomic offset and habitat suitability in assessing invasion risks under climate change. Moreover, incorporating adaptive genomic variation into predictive models can significantly enhance future invasion risk predictions and enable effective management strategies for biological invasions in the future.

Predicting the future of our oceans-Evaluating genomic forecasting approaches in marine species

Climate change is restructuring biodiversity on multiple scales and there is a pressing need to understand the downstream ecological and genomic consequences of this change. Recent advancements in the field of eco-evolutionary genomics have sought to include evolutionary processes in forecasting species' responses to climate change (e.g., genomic offset), but to date, much of this work has focused on terrestrial species. Coastal and offshore species, and the fisheries they support, may be even more vulnerable to climate change than their terrestrial counterparts, warranting a critical appraisal of these approaches in marine systems. First, we synthesize knowledge about the genomic basis of adaptation in marine species, and then we discuss the few examples where genomic forecasting has been applied in marine systems. Next, we identify the key challenges in validating genomic offset estimates in marine species, and we advocate for the inclusion of historical sampling data and hindcasting in the validation phase. Lastly, we describe a workflow to guide marine managers in incorporating these predictions into the decision-making process.

How Can Genomics Help or Hinder Wildlife Conservation?

Genomic data are becoming increasingly affordable and easy to collect, and new tools for their analysis are appearing rapidly. Conservation biologists are interested in using this information to assist in management and planning but are typically limited financially and by the lack of genomic resources available for non-model taxa. It is therefore important to be aware of the pitfalls as well as the benefits of applying genomic approaches. Here, we highlight recent methods aimed at standardizing population assessments of genetic variation, inbreeding, and forms of genetic load and methods that help identify past and ongoing patterns of genetic interchange between populations, including those subjected to recent disturbance. We emphasize challenges in applying some of these methods and the need for adequate bioinformatic support. We also consider the promises and challenges of applying genomic approaches to understand adaptive changes in natural populations to predict their future adaptive capacity.

How Fish Population Genomics Can Promote Sustainable Fisheries: A Road Map

Maintenance of genetic diversity in marine fishes targeted by commercial fishing is a grand challenge for the future. Most of these species are abundant and therefore important for marine ecosystems and food security. Here, we present a road map of how population genomics can promote sustainable fisheries. In these species, the development of reference genomes and whole genome sequencing is key, because genetic differentiation at neutral loci is usually low due to large population sizes and gene flow. First, baseline allele frequencies representing genetically differentiated populations within species must be established. These can then be used to accurately determine the composition of mixed samples, forming the basis for population demographic analysis to inform sustainably set fish quotas. SNP-chip analysis is a cost-effective method for determining baseline allele frequencies and for population identification in mixed samples. Finally, we describe how genetic marker analysis can transform stock identification and management.

Seascapes Shaped the Local Adaptation and Population Structure of South China Coast Yellowfin Seabream (Acanthopagrus latus)

Understanding the genetic composition and regional adaptation of marine species under environmental heterogeneity and fishing pressure is crucial for responsible management. In order to understand the genetic diversity and adaptability of yellowfin seabream (Acanthopagrus latus) along southern China coast, this study was conducted a seascape genome analysis on yellowfin seabream from the ecologically diverse coast, spanning over 1600 km. A total of 92 yellowfin seabream individuals from 15 sites were performed whole-genome resequencing, and 4,383,564 high-quality single nucleotide polymorphisms (SNPs) were called. By conducting a genotype-environment association analysis, 29,951 adaptive and 4,328,299 neutral SNPs were identified. The yellowfin seabream exhibited two distinct population structures, despite high gene flow between sites. The seascape genome analysis revealed that genetic structure was influenced by a variety of factors including salinity gradients, habitat distance, and ocean currents. The frequency of allelic variation at the candidate loci changed with the salinity gradient. Annotation of these loci revealed that most of the genes are associated with osmoregulation, such as kcnab2a, kcnk5a, and slc47a1. These genes are significantly enriched in pathways associated with ion transport including G protein-coupled receptor activity, transmembrane signaling receptor activity, and transporter activity. Overall, our findings provide insights into how seascape heterogeneity affects adaptive evolution, while providing important information for regional management in yellowfin seabream populations.

Disease-driven top predator decline affects mesopredator population genomic structure

Top predator declines are pervasive and often have dramatic effects on ecological communities via changes in food web dynamics, but their evolutionary consequences are virtually unknown. Tasmania's top terrestrial predator, the Tasmanian devil, is declining due to a lethal transmissible cancer. Spotted-tailed quolls benefit via mesopredator release, and they alter their behaviour and resource use concomitant with devil declines and increased disease duration. Here, using a landscape community genomics framework to identify environmental drivers of population genomic structure and signatures of selection, we show that these biotic factors are consistently among the top variables explaining genomic structure of the quoll. Landscape resistance negatively correlates with devil density, suggesting that devil declines will increase quoll genetic subdivision over time, despite no change in quoll densities detected by camera trap studies. Devil density also contributes to signatures of selection in the quoll genome, including genes associated with muscle development and locomotion. Our results provide some of the first evidence of the evolutionary impacts of competition between a top predator and a mesopredator species in the context of a trophic cascade. As top predator declines are increasing globally, our framework can serve as a model for future studies of evolutionary impacts of altered ecological interactions.

Projection of current and future distribution of adaptive genetic units in an alpine ungulate

Climate projections predict major changes in alpine environments by the end of the 21st century. To avoid climate-induced maladaptation and extinction, many animal populations will either need to move to more suitable habitats or adapt in situ to novel conditions. Since populations of a species exhibit genetic variation related to local adaptation, it is important to incorporate this variation into predictive models to help assess the ability of the species to survive climate change. Here, we evaluate how the adaptive genetic variation of a mountain ungulate-the Northern chamois (Rupicapra rupicapra)-could be impacted by future global warming. Based on genotype-environment association analyses of 429 chamois using a ddRAD sequencing approach, we identified genetic variation associated with climatic gradients across the European Alps. We then delineated adaptive genetic units and projected the optimal distribution of these adaptive groups in the future. Our results suggest the presence of local adaptation to climate in Northern chamois with similar genetic adaptive responses in geographically distant but climatically similar populations. Furthermore, our results predict that future climatic changes will modify the Northern chamois adaptive landscape considerably, with various degrees of maladaptation risk.

Space-for-time substitutions in climate change ecology and evolution

In an epoch of rapid environmental change, understanding and predicting how biodiversity will respond to a changing climate is an urgent challenge. Since we seldom have sufficient long-term biological data to use the past to anticipate the future, spatial climate-biotic relationships are often used as a proxy for predicting biotic responses to climate change over time. These 'space-for-time substitutions' (SFTS) have become near ubiquitous in global change biology, but with different subfields largely developing methods in isolation. We review how climate-focussed SFTS are used in four subfields of ecology and evolution, each focussed on a different type of biotic variable - population phenotypes, population genotypes, species' distributions, and ecological communities. We then examine the similarities and differences between subfields in terms of methods, limitations and opportunities. While SFTS are used for a wide range of applications, two main approaches are applied across the four subfields: spatial in situ gradient methods and transplant experiments. We find that SFTS methods share common limitations relating to (i) the causality of identified spatial climate-biotic relationships and (ii) the transferability of these relationships, i.e. whether climate-biotic relationships observed over space are equivalent to those occurring over time. Moreover, despite widespread application of SFTS in climate change research, key assumptions remain largely untested. We highlight opportunities to enhance the robustness of SFTS by addressing key assumptions and limitations, with a particular emphasis on where approaches could be shared between the four subfields.

The meaning of wild: Genetic and adaptive consequences from large-scale releases of domestic mallards

The translocation of individuals around the world is leading to rising incidences of anthropogenic hybridization, particularly between domestic and wild congeners. We apply a landscape genomics approach for thousands of mallard (Anas platyrhynchos) samples across continental and island populations to determine the result of over a century of supplementation practices. We establish that a single domestic game-farm mallard breed is the source for contemporary release programs in Eurasia and North America, as well as for established feral populations in New Zealand and Hawaii. In particular, we identify central Europe and eastern North America as epicenters of ongoing anthropogenic hybridization, and conclude that the release of game-farm mallards continues to affect the genetic integrity of wild mallards. Conversely, self-sustaining feral populations in New Zealand and Hawaii not only show strong differentiation from their original stock, but also signatures of local adaptation occurring in less than a half-century since game-farm mallard releases have ceased. We conclude that 'wild' is not singular, and that even feral populations are capable of responding to natural processes. Although considered paradoxical to biological conservation, understanding the capacity for wildness among feral and feral admixed populations in human landscapes is critical as such interactions increase in the Anthropocene.

Re-thinking the environment in landscape genomics

Detecting the extrinsic selective pressures shaping genomic variation is critical for a better understanding of adaptation and for forecasting evolutionary responses of natural populations to changing environmental conditions. With increasing availability of geo-referenced environmental data, landscape genomics provides unprecedented insights into how genomic variation and underlying gene functions affect traits potentially under selection. Yet, the robustness of genotype-environment associations used in landscape genomics remains tempered due to various limitations, including the characteristics of environmental data used, sampling designs employed, and statistical frameworks applied. Here, we argue that using complementary or new environmental data sources and well-informed sampling designs may help improve the detection of selective pressures underlying patterns of local adaptation in various organisms and environments.

Understanding local plant extinctions before it is too late: bridging evolutionary genomics with global ecology

Understanding evolutionary genomic and population processes within a species range is key to anticipating the extinction of plant species before it is too late. However, most models of biodiversity risk under global change do not account for the genetic variation and local adaptation of different populations. Population diversity is critical to understanding extinction because different populations may be more or less susceptible to global change and, if lost, would reduce the total diversity within a species. Two new modeling frameworks advance our understanding of extinction from a population and evolutionary angle: Rapid climate change-driven disruptions in population adaptation are predicted from associations between genomes and local climates. Furthermore, losses of population diversity from global land-use transformations are estimated by scaling relationships of species' genomic diversity with habitat area. Overall, these global eco-evolutionary methods advance the predictability - and possibly the preventability - of the ongoing extinction of plant species.

From common gardens to candidate genes: exploring local adaptation to climate in red spruce

Local adaptation to climate is common in plant species and has been studied in a range of contexts, from improving crop yields to predicting population maladaptation to future conditions. The genomic era has brought new tools to study this process, which was historically explored through common garden experiments. In this study, we combine genomic methods and common gardens to investigate local adaptation in red spruce and identify environmental gradients and loci involved in climate adaptation. We first use climate transfer functions to estimate the impact of climate change on seedling performance in three common gardens. We then explore the use of multivariate gene-environment association methods to identify genes underlying climate adaptation, with particular attention to the implications of conducting genome scans with and without correction for neutral population structure. This integrative approach uncovered phenotypic evidence of local adaptation to climate and identified a set of putatively adaptive genes, some of which are involved in three main adaptive pathways found in other temperate and boreal coniferous species: drought tolerance, cold hardiness, and phenology. These putatively adaptive genes segregated into two 'modules' associated with different environmental gradients. This study nicely exemplifies the multivariate dimension of adaptation to climate in trees.

Climate-induced range shifts drive adaptive response via spatio-temporal sieving of alleles

Quaternary climate fluctuations drove many species to shift their geographic ranges, in turn shaping their genetic structures. Recently, it has been argued that adaptation may have accompanied species range shifts via the "sieving" of genotypes during colonisation and establishment. However, this has not been directly demonstrated, and knowledge remains limited on how different evolutionary forces, which are typically investigated separately, interacted to jointly mediate species responses to past climatic change. Here, through whole-genome re-sequencing of over 1200 individuals of the carnation Dianthus sylvestris coupled with integrated population genomic and gene-environment models, we reconstruct the past neutral and adaptive landscape of this species as it was shaped by the Quaternary glacial cycles. We show that adaptive responses emerged concomitantly with the post-glacial range shifts and expansions of this species in the last 20 thousand years. This was due to the heterogenous sieving of adaptive alleles across space and time, as populations expanded out of restrictive glacial refugia into the broader and more heterogeneous range of habitats available in the present-day inter-glacial. Our findings reveal a tightly-linked interplay of migration and adaptation under past climate-induced range shifts, which we show is key to understanding the spatial patterns of adaptive variation we see in species today.

Genotype-environment associations to reveal the molecular basis of environmental adaptation

A fundamental goal in plant biology is to identify and understand the variation underlying plants' adaptation to their environment. Climate change has given new urgency to this goal, as society aims to accelerate adaptation of ecologically important plant species, endangered plant species, and crops to hotter, less predictable climates. In the pre-genomic era, identifying adaptive alleles was painstaking work, leveraging genetics, molecular biology, physiology, and ecology. Now, the rise of genomics and new computational approaches may facilitate this research. Genotype-environment associations (GEAs) use statistical associations between allele frequency and environment of origin to test the hypothesis that allelic variation at a given gene is adapted to local environments. Researchers may scan the genome for GEAs to generate hypotheses on adaptive genetic variants (environmental genome-wide association studies). Despite the rapid adoption of these methods, many important questions remain about the interpretation of GEA findings, which arise from fundamental unanswered questions on the genetic architecture of adaptation and limitations inherent to association-based analyses. We outline strategies to ground GEAs in the underlying hypotheses of genetic architecture and better test GEA-generated hypotheses using genetics and ecophysiology. We provide recommendations for new users who seek to learn about the molecular basis of adaptation. When combined with a rigorous hypothesis testing framework, GEAs may facilitate our understanding of the molecular basis of climate adaptation for plant improvement.

Using landscape genomics to assess local adaptation and genomic vulnerability of a perennial herb Tetrastigma hemsleyanum (Vitaceae) in subtropical China

Understanding adaptive genetic variation of plant populations and their vulnerabilities to climate change are critical to preserve biodiversity and subsequent management interventions. To this end, landscape genomics may represent a cost-efficient approach for investigating molecular signatures underlying local adaptation. Tetrastigma hemsleyanum is, in its native habitat, a widespread perennial herb of warm-temperate evergreen forest in subtropical China. Its ecological and medicinal values constitute a significant revenue for local human populations and ecosystem. Using 30,252 single nucleotide polymorphisms (SNPs) derived from reduced-representation genome sequencing in 156 samples from 24 sites, we conducted a landscape genomics study of the T. hemsleyanum to elucidate its genomic variation across multiple climate gradients and genomic vulnerability to future climate change. Multivariate methods identified that climatic variation explained more genomic variation than that of geographical distance, which implied that local adaptation to heterogeneous environment might represent an important source of genomic variation. Among these climate variables, winter precipitation was the strongest predictor of the contemporary genetic structure. F _ST outlier tests and environment association analysis totally identified 275 candidate adaptive SNPs along the genetic and environmental gradients. SNP annotations of these putatively adaptive loci uncovered gene functions associated with modulating flowering time and regulating plant response to abiotic stresses, which have implications for breeding and other special agricultural aims on the basis of these selection signatures. Critically, modelling revealed that the high genomic vulnerability of our focal species via a mismatch between current and future genotype-environment relationships located in central-northern region of the T. hemsleyanum's range, where populations require proactive management efforts such as assistant adaptation to cope with ongoing climate change. Taken together, our results provide robust evidence of local climate adaption for T. hemsleyanum and further deepen our understanding of adaptation basis of herbs in subtropical China.

Cell-type-specific co-expression inference from single cell RNA-sequencing data

The inference of gene co-expressions from microarray and RNA-sequencing data has led to rich insights on biological processes and disease mechanisms. However, the bulk samples analyzed in most studies are a mixture of different cell types. As a result, the inferred co-expressions are confounded by varying cell type compositions across samples and only offer an aggregated view of gene regulations that may be distinct across different cell types. The advancement of single cell RNA-sequencing (scRNA-seq) technology has enabled the direct inference of co-expressions in specific cell types, facilitating our understanding of cell-type-specific biological functions. However, the high sequencing depth variations and measurement errors in scRNA-seq data present significant challenges in inferring cell-type-specific gene co-expressions, and these issues have not been adequately addressed in the existing methods. We propose a statistical approach, CS-CORE, for estimating and testing cell-type-specific co-expressions, built on a general expression-measurement model that explicitly accounts for sequencing depth variations and measurement errors in the observed single cell data. Systematic evaluations show that most existing methods suffer from inflated false positives and biased co-expression estimates and clustering analysis, whereas CS-CORE has appropriate false positive control, unbiased co-expression estimates, good statistical power and satisfactory performance in downstream co-expression analysis. When applied to analyze scRNA-seq data from postmortem brain samples from Alzheimer’s disease patients and controls and blood samples from COVID-19 patients and controls, CS-CORE identified cell-type-specific co-expressions and differential co-expressions that were more reproducible and/or more enriched for relevant biological pathways than those inferred from other methods.

Local adaptation: Causal agents of selection and adaptive trait divergence

Divergent selection across the landscape can favor the evolution of local adaptation in populations experiencing contrasting conditions. Local adaptation is widely observed in a diversity of taxa, yet we have a surprisingly limited understanding of the mechanisms that give rise to it. For instance, few have experimentally confirmed the biotic and abiotic variables that promote local adaptation, and fewer yet have identified the phenotypic targets of selection that mediate local adaptation. Here, we highlight critical gaps in our understanding of the process of local adaptation and discuss insights emerging from in-depth investigations of the agents of selection that drive local adaptation, the phenotypes they target, and the genetic basis of these phenotypes. We review historical and contemporary methods for assessing local adaptation, explore whether local adaptation manifests differently across life history, and evaluate constraints on local adaptation.

Genomic insights into local adaptation and future climate-induced vulnerability of a keystone forest tree in East Asia

Rapid global climate change is posing a substantial threat to biodiversity. The assessment of population vulnerability and adaptive capacity under climate change is crucial for informing conservation and mitigation strategies. Here we generate a chromosome-scale genome assembly and re-sequence genomes of 230 individuals collected from 24 populations for Populus koreana, a pioneer and keystone tree species in temperate forests of East Asia. We integrate population genomics and environmental variables to reveal a set of climate-associated single-nucleotide polymorphisms, insertion/deletions and structural variations, especially numerous adaptive non-coding variants distributed across the genome. We incorporate these variants into an environmental modeling scheme to predict a highly spatiotemporal shift of this species in response to future climate change. We further identify the most vulnerable populations that need conservation priority and many candidate genes and variants that may be useful for forest tree breeding with special aims. Our findings highlight the importance of integrating genomic and environmental data to predict adaptive capacity of a key forest to rapid climate change in the future.

Eco-evolutionary causes and consequences of rarity in plants: a meta-analysis

Species differ dramatically in their prevalence in the natural world, with many species characterized as rare due to restricted geographic distribution, low local abundance and/or habitat specialization. We investigated the ecoevolutionary causes and consequences of rarity with phylogenetically controlled metaanalyses of population genetic diversity, fitness and functional traits in rare and common congeneric plant species. Our syntheses included 252 rare species and 267 common congeners reported in 153 peer-reviewed articles published from 1978 to 2020 and one manuscript in press. Rare species have reduced population genetic diversity, depressed fitness and smaller reproductive structures than common congeners. Rare species also could suffer from inbreeding depression and reduced fertilization efficiency. By limiting their capacity to adapt and migrate, these characteristics could influence contemporary patterns of rarity and increase the susceptibility of rare species to rapid environmental change. We recommend that future studies present more nuanced data on the extent of rarity in focal species, expose rare and common species to ecologically relevant treatments, including reciprocal transplants, and conduct quantitative genetic and population genomic analyses across a greater array of systems. This research could elucidate the processes that contribute to rarity and generate robust predictions of extinction risks under global change.

Climate Adaptation, Drought Susceptibility, and Genomic-Informed Predictions of Future Climate Refugia for the Australian Forest Tree Eucalyptus globulus

Understanding the capacity of forest tree species to adapt to climate change is of increasing importance for managing forest genetic resources. Through a genomics approach, we modelled spatial variation in climate adaptation within the Australian temperate forest tree Eucalyptus globulus, identified putative climate drivers of this genomic variation, and predicted locations of future climate refugia and populations at-risk of future maladaptation. Using 812,158 SNPs across 130 individuals from 30 populations (i.e., localities) spanning the species’ natural range, a gradientForest algorithm found 1177 SNPs associated with locality variation in home-site climate (climate-SNPs), putatively linking them to climate adaptation. Very few climate-SNPs were associated with population-level variation in drought susceptibility, signalling the multi-faceted nature and complexity of climate adaptation. Redundancy analysis (RDA) showed 24% of the climate-SNP variation could be explained by annual precipitation, isothermality, and maximum temperature of the warmest month. Spatial predictions of the RDA climate vectors associated with climate-SNPs allowed mapping of genomically informed climate selective surfaces across the species’ range under contemporary and projected future climates. These surfaces suggest over 50% of the current distribution of E. globulus will be outside the modelled adaptive range by 2070 and at risk of climate maladaptation. Such surfaces present a new integrated approach for natural resource managers to capture adaptive genetic variation and plan translocations in the face of climate change.

Differential Adaptive Potential and Vulnerability to Climate-Driven Habitat Loss in Brazilian Mangroves

Geographic and environmental differences have been identified as factors influencing Brazilian mangrove trees' genetic diversity. Geographically, distinct species have convergent spatial genetic structures, indicating a limited gene flow between northern and southern populations. Environmentally, genomic studies and common garden experiments have found evidence of local adaptations along the latitudinal gradient of the Brazilian coast. However, little is known about how such adaptive heterogeneity could be affected by a rapidly changing climate in the coming decades, and the combination of deforestation and climate-induced habitat loss may affect these forests and their genetic diversity. Here, we applied two genomic-environmental association methods to model the turnover of potentially adaptive alleles for two dominant mangrove trees: Avicennia germinans and A. schaueriana. We analyzed a total of 134 individuals from six populations of A. germinans and 10 populations of A. schaueriana spanning the Brazilian coast from 1 °S to 28 °S. Gradient forest models identified temperature-related variables as the most important predictors for A. germinans outlier loci, whereas both temperature and precipitation were important for A. schaueriana. We modeled allele frequencies and projected them for future climatic scenarios to estimate adaptively driven vulnerability. We assessed climate-driven habitat loss through climate-only distribution models and calculated annual deforestation rates for each sampled region. Finally, to assess the vulnerability of individual populations, we combined the environmental suitability, deforestation data, and adaptive vulnerability projections. For both species, subtropical populations presented a higher vulnerability than equatorial populations to climate-driven habitat loss. We also identified deforestation rates at the sampled sites that were alarmingly higher than the global average mangrove deforestation rate. Our results provide improved estimates of the impacts of ongoing climate change and human-caused habitat loss on the distribution of mangroves and highlight the importance of site-based conservation strategies that consider individual subtropical and equatorial mangrove forests.

Seeing the Forest for the trees: Assessing genetic offset predictions from Gradient Forest

Gradient Forest (GF) is a machine learning algorithm designed to analyze spatial patterns of biodiversity as a function of environmental gradients. An offset measure between the GF‐predicted environmental association of adapted alleles and a new environment (GF Offset) is increasingly being used to predict the loss of environmentally adapted alleles under rapid environmental change, but remains mostly untested for this purpose. Here, we explore the robustness of GF Offset to assumption violations, and its relationship to measures of fitness, using SLiM simulations with explicit genome architecture and a spatial metapopulation. We evaluate measures of GF Offset in: (1) a neutral model with no environmental adaptation; (2) a monogenic “population genetic” model with a single environmentally adapted locus; and (3) a polygenic “quantitative genetic” model with two adaptive traits, each adapting to a different environment. We found GF Offset to be broadly correlated with fitness offsets under both single locus and polygenic architectures. However, neutral demography, genomic architecture, and the nature of the adaptive environment can all confound relationships between GF Offset and fitness. GF Offset is a promising tool, but it is important to understand its limitations and underlying assumptions, especially when used in the context of predicting maladaptation.

Genomics and adaptation in forest ecosystems

Rapid human-induced environmental changes like climate warming represent a challenge for forest ecosystems. Due to their biological complexity and the long generation time of their keystone tree species, genetic adaptation in these ecosystems might not be fast enough to keep track with conditions changing at such a fast pace. The study of adaptation to environmental change and its genetic mechanisms is therefore key for ensuring a sustainable support and management of forests. The 4-day conference of the European Research Group EvolTree (https://www.evoltree.eu) on the topic of "Genomics and Adaptation in Forest Ecosystems" brought together over 130 scientists to present and discuss the latest developments and findings in forest evolutionary research. Genomic studies in forest trees have long been hampered by the lack of high-quality genomics resources and affordable genotyping methods. This has dramatically changed in the last few years; the conference impressively showed how such tools are now being applied to study past demography, adaptation and interactions with associated organisms. Moreover, genomic studies are now finally also entering the world of conservation and forest management, for example by measuring the value or cost of interspecific hybridization and introgression, assessing the vulnerability of species and populations to future change, or accurately delineating evolutionary significant units. The newly launched conference series of EvolTree will hopefully play a key role in the exchange and synthesis of such important investigations.

SUPPLEMENTARY INFORMATION

The online version contains supplementary material available at 10.1007/s11295-022-01542-1.

Landscape Genomics in Tree Conservation Under a Changing Environment

Understanding the genetic basis of how species respond to changing environments is essential to the conservation of species. However, the molecular mechanisms of adaptation remain largely unknown for long-lived tree species which always have large population sizes, long generation time, and extensive gene flow. Recent advances in landscape genomics can reveal the signals of adaptive selection linking genetic variations and landscape characteristics and therefore have created novel insights into tree conservation strategies. In this review article, we first summarized the methods of landscape genomics used in tree conservation and elucidated the advantages and disadvantages of these methods. We then highlighted the newly developed method "Risk of Non-adaptedness," which can predict the genetic offset or genomic vulnerability of species via allele frequency change under multiple scenarios of climate change. Finally, we provided prospects concerning how our introduced approaches of landscape genomics can assist policymaking and improve the existing conservation strategies for tree species under the ongoing global changes.

Genomic and common garden approaches yield complementary results for quantifying environmental drivers of local adaptation in rubber rabbitbrush, a foundational Great Basin shrub

The spatial structure of genomic and phenotypic variation across populations reflects historical and demographic processes as well as evolution via natural selection. Characterizing such variation can provide an important perspective for understanding the evolutionary consequences of changing climate and for guiding ecological restoration. While evidence for local adaptation has been traditionally evaluated using phenotypic data, modern methods for generating and analyzing landscape genomic data can directly quantify local adaptation by associating allelic variation with environmental variation. Here, we analyze both genomic and phenotypic variation of rubber rabbitbrush (Ericameria nauseosa), a foundational shrub species of western North America. To quantify landscape genomic structure and provide perspective on patterns of local adaptation, we generated reduced representation sequencing data for 17 wild populations (222 individuals; 38,615 loci) spanning a range of environmental conditions. Population genetic analyses illustrated pronounced landscape genomic structure jointly shaped by geography and environment. Genetic-environment association (GEA) analyses using both redundancy analysis (RDA) and a machine-learning approach (Gradient Forest) indicated environmental variables (precipitation seasonality, slope, aspect, elevation, and annual precipitation) influenced spatial genomic structure and were correlated with allele frequency shifts indicative of local adaptation at a consistent set of genomic regions. We compared our GEA-based inference of local adaptation with phenotypic data collected by growing seeds from each population in a greenhouse common garden. Population differentiation in seed weight, emergence, and seedling traits was associated with environmental variables (e.g., precipitation seasonality) that were also implicated in GEA analyses, suggesting complementary conclusions about the drivers of local adaptation across different methods and data sources. Our results provide a baseline understanding of spatial genomic structure for E. nauseosa across the western Great Basin and illustrate the utility of GEA analyses for detecting the environmental causes and genetic signatures of local adaptation in a widely distributed plant species of restoration significance.

The evolutionary genomics of species' responses to climate change

Climate change is a threat to biodiversity. One way that this threat manifests is through pronounced shifts in the geographical range of species over time. To predict these shifts, researchers have primarily used species distribution models. However, these models are based on assumptions of niche conservatism and do not consider evolutionary processes, potentially limiting their accuracy and value. To incorporate evolution into the prediction of species' responses to climate change, researchers have turned to landscape genomic data and examined information about local genetic adaptation using climate models. Although this is an important advancement, this approach currently does not include other evolutionary processes-such as gene flow, population dispersal and genomic load-that are critical for predicting the fate of species across the landscape. Here, we briefly review the current practices for the use of species distribution models and for incorporating local adaptation. We next discuss the rationale and theory for considering additional processes, reviewing how they can be incorporated into studies of species' responses to climate change. We summarize with a conceptual framework of how manifold layers of information can be combined to predict the potential response of specific populations to climate change. We illustrate all of the topics using an exemplar dataset and provide the source code as potential tutorials. This Perspective is intended to be a step towards a more comprehensive integration of population genomics with climate change science.

Prospects and limitations of genomic offset in conservation management

In nature conservation, there is keen interest in predicting how populations will respond to environmental changes such as climate change. These predictions can help determine whether a population can be self-sustaining under future alterations of its habitat or whether it may require human intervention such as protection, restoration, or assisted migration. An increasingly popular approach in this respect is the concept of genomic offset, which combines genomic and environmental data from different time points and/or locations to assess the degree of possible maladaptation to new environmental conditions. Here, we argue that the concept of genomic offset holds great potential, but an exploration of its risks and limitations is needed to use it for recommendations in conservation or assisted migration. After briefly describing the concept, we list important issues to consider (e.g., statistical frameworks, population genetic structure, migration, independent evidence) when using genomic offset or developing these methods further. We conclude that genomic offset is an area of development that still lacks some important features and should be used in combination with other approaches to inform conservation measures.