Machine learning identifies specific habitats associated with genetic connectivity in Hyla squirella

Predicting Environmental and Ecological Drivers of Human Population Structure

Landscape, climate, and culture can all structure human populations, but few existing methods are designed to simultaneously disentangle among a large number of variables in explaining genetic patterns. We developed a machine learning method for identifying the variables which best explain migration rates, as measured by the coalescent-based program MAPS that uses shared identical by descent tracts to infer spatial migration across a region of interest. We applied our method to 30 human populations in eastern Africa with high-density single nucleotide polymorphism array data. The remarkable diversity of ethnicities, languages, and environments in this region offers a unique opportunity to explore the variables that shape migration and genetic structure. We explored more than 20 spatial variables relating to landscape, climate, and presence of tsetse flies. The full model explained ∼40% of the variance in migration rate over the past 56 generations. Precipitation, minimum temperature of the coldest month, and elevation were the variables with the highest impact. Among the three groups of tsetse flies, the most impactful was fusca which transmits livestock trypanosomiasis. We also tested for adaptation to high elevation among Ethiopian populations. We did not identify well-known genes related to high elevation, but we did find signatures of positive selection related to metabolism and disease. We conclude that the environment has influenced the migration and adaptation of human populations in eastern Africa; the remaining variance in structure is likely due in part to cultural or other factors not captured in our model.

Estimating resistance surfaces using gradient forest and allelic frequencies

Understanding landscape connectivity has become a global priority for mitigating the impact of landscape fragmentation on biodiversity. Connectivity methods that use link-based methods traditionally rely on relating pairwise genetic distance between individuals or demes to their landscape distance (e.g., geographic distance, cost distance). In this study, we present an alternative to conventional statistical approaches to refine cost surfaces by adapting the gradient forest approach to produce a resistance surface. Used in community ecology, gradient forest is an extension of random forest, and has been implemented in genomic studies to model species genetic offset under future climatic scenarios. By design, this adapted method, resGF, has the ability to handle multiple environmental predicators and is not subjected to traditional assumptions of linear models such as independence, normality and linearity. Using genetic simulations, resistance Gradient Forest (resGF) performance was compared to other published methods (maximum likelihood population effects model, random forest-based least-cost transect analysis and species distribution model). In univariate scenarios, resGF was able to distinguish the true surface contributing to genetic diversity among competing surfaces better than the compared methods. In multivariate scenarios, the gradient forest approach performed similarly to the other random forest-based approach using least-cost transect analysis but outperformed MLPE-based methods. Additionally, two worked examples are provided using two previously published data sets. This machine learning algorithm has the potential to improve our understanding of landscape connectivity and inform long-term biodiversity conservation strategies.

Landscape genetics of a sub-alpine toad: climate change predicted to induce upward range shifts via asymmetrical migration corridors

Climate change is expected to have a major hydrological impact on the core breeding habitat and migration corridors of many amphibians in the twenty-first century. The Yosemite toad (Anaxyrus canorus) is a species of meadow-specializing amphibian endemic to the high-elevation Sierra Nevada Mountains of California. Despite living entirely on federal lands, it has recently faced severe extirpations, yet our understanding of climatic influences on population connectivity is limited. In this study, we used a previously published double-digest RADseq dataset along with numerous remotely sensed habitat features in a landscape genetics framework to answer two primary questions in Yosemite National Park: (1) Which fine-scale climate, topographic, soil, and vegetation features most facilitate meadow connectivity? (2) How is climate change predicted to influence both the magnitude and net asymmetry of genetic migration? We developed an approach for simultaneously modeling multiple toad migration paths, akin to circuit theory, except raw environmental features can be separately considered. Our workflow identified the most likely migration corridors between meadows and used the unique cubist machine learning approach to fit and forecast environmental models of connectivity. We identified the permuted modeling importance of numerous snowpack-related features, such as runoff and groundwater recharge. Our results highlight the importance of considering phylogeographic structure, and asymmetrical migration in landscape genetics. We predict an upward elevational shift for this already high-elevation species, as measured by the net vector of anticipated genetic movement, and a north-eastward shift in species distribution via the network of genetic migration corridors across the park.

An ensemble learning approach to map the genetic connectivity of the parasitoid Stethynium empoasca (Hymenoptera: Mymaridae) and identify the key influencing environmental and landscape factors

The effect of landscape patterns and environmental factors on the population structure and genetic diversity of organisms is well-documented. However, this effect is still unclear in the case of Mymaridae parasitoids. Despite recent advances in machine learning methods for landscape genetics, ensemble learning still needs further investigation. Here, we evaluated the performance of different boosting algorithms and analyzed the effects of landscape and environmental factors on the genetic variations in the tea green leafhopper parasitoid Stethynium empoasca (Hymenoptera: Mymaridae). The S. empoasca populations showed a distinct pattern of isolation by distance. The minimum temperature of the coldest month, annual precipitation, the coverage of evergreen/deciduous needleleaf trees per 1 km2, and the minimum precipitation of the warmest quarter were identified as the dominant factors affecting the genetic divergence of S. empoasca populations. Notably, compared to previous machine learning studies, our model showed an unprecedented accuracy (r = 0.87) for the prediction of genetic differentiation. These findings not only demonstrated how the landscape shaped S. empoasca genetics but also provided an essential basis for developing conservation strategies for this biocontrol agent. In a broader sense, this study demonstrated the importance and efficiency of ensemble learning in landscape genetics.

Location and Species Matters: Variable Influence of the Environment on the Gene Flow of Imperiled, Native and Invasive Cottontails

The environment plays an important role in the movement of individuals and their associated genes among populations, which facilitates gene flow. Gene flow can help maintain the genetic diversity both within and between populations and counter the negative impact of genetic drift, which can decrease the fitness of individuals. Sympatric species can have different habitat preferences, and thus can exhibit different patterns of genetic variability and population structure. The specialist-generalist variation hypothesis (SGVH) predicts that specialists will have lower genetic diversity, lower effective population sizes (Ne), and less gene flow among populations. In this study, we used spatially explicit, individual-based comparative approaches to test SGVH predictions in two sympatric cottontail species and identify environmental variables that influence their gene flow. New England cottontail (Sylvilagus transitionalis) is the only native cottontail in the Northeast US, an early successional habitat specialist, and a species of conservation concern. Eastern cottontail (S. floridanus) is an invasive species in the Northeast US and a habitat generalist. We characterized each species' genomic variation by developing double-digest Restriction-site Associated DNA sequence single nucleotide polymorphism markers, quantified their habitat with Geographic Information System environmental variables, and conducted our analyses at multiple scales. Surprisingly, both species had similar levels of genetic diversity and eastern cottontail's Ne was only higher than New England cottontail in one of three subregions. At a regional level, the population clusters of New England cottontail were more distinct than eastern cottontail, but the subregional levels showed more geographic areas of restricted gene flow for eastern cottontail than New England cottontail. In general, the environmental variables had the predicted effect on each species' gene flow. However, the most important environmental variable varied by subregion and species, which shows that location and species matter. Our results provide partial support for the SGVH and the identification of environmental variables that facilitate or impede gene flow can be used to help inform management decisions to conserve New England cottontail.

A machine-learning approach to map landscape connectivity in Aedes aegypti with genetic and environmental data

Aedes mosquitoes are projected to continue expanding their ranges, which could expose millions more humans to the diseases they carry. The implementation of vector control methods ranging from traditional (e.g., insecticides) to cutting edge (e.g., genetic modification) could be improved with landscape connectivity maps and increased understanding of the factors that affect mosquito dispersal. Here we present an iterative random forest method for integrating genetic and environmental data to map landscape connectivity. We achieve a correlation of 0.83 between the model’s predicted genetic distance and actual genetic distance. We produce a genetic connectivity map for the southern tier of the United States and discuss important factors to consider in mosquito control, e.g., the release of genetically modified mosquitoes.

Mapping landscape connectivity is important for controlling invasive species and disease vectors. Current landscape genetics methods are often constrained by the subjectivity of creating resistance surfaces and the difficulty of working with interacting and correlated environmental variables. To overcome these constraints, we combine the advantages of a machine-learning framework and an iterative optimization process to develop a method for integrating genetic and environmental (e.g., climate, land cover, human infrastructure) data. We validate and demonstrate this method for the Aedes aegypti mosquito, an invasive species and the primary vector of dengue, yellow fever, chikungunya, and Zika. We test two contrasting metrics to approximate genetic distance and find Cavalli-Sforza–Edwards distance (CSE) performs better than linearized F_ST. The correlation (R) between the model’s predicted genetic distance and actual distance is 0.83. We produce a map of genetic connectivity for Ae. aegypti’s range in North America and discuss which environmental and anthropogenic variables are most important for predicting gene flow, especially in the context of vector control.

A machine learning approach to integrating genetic and ecological data in tsetse flies (Glossina pallidipes) for spatially explicit vector control planning

Vector control is an effective strategy for reducing vector-borne disease transmission, but requires knowledge of vector habitat use and dispersal patterns. Our goal was to improve this knowledge for the tsetse species Glossina pallidipes, a vector of human and animal African trypanosomiasis, which are diseases that pose serious health and socioeconomic burdens across sub-Saharan Africa. We used random forest regression to (i) build and integrate models of G. pallidipes habitat suitability and genetic connectivity across Kenya and northern Tanzania and (ii) provide novel vector control recommendations. Inputs for the models included field survey records from 349 trap locations, genetic data from 11 microsatellite loci from 659 flies and 29 sampling sites, and remotely sensed environmental data. The suitability and connectivity models explained approximately 80% and 67% of the variance in the occurrence and genetic data and exhibited high accuracy based on cross-validation. The bivariate map showed that suitability and connectivity vary independently across the landscape and was used to inform our vector control recommendations. Post hoc analyses show spatial variation in the correlations between the most important environmental predictors from our models and each response variable (e.g., suitability and connectivity) as well as heterogeneity in expected future climatic change of these predictors. The bivariate map suggests that vector control is most likely to be successful in the Lake Victoria Basin and supports the previous recommendation that G. pallidipes from most of eastern Kenya should be managed as a single unit. We further recommend that future monitoring efforts should focus on tracking potential changes in vector presence and dispersal around the Serengeti and the Lake Victoria Basin based on projected local climatic shifts. The strong performance of the spatial models suggests potential for our integrative methodology to be used to understand future impacts of climate change in this and other vector systems.

Cryptic chytridiomycosis linked to climate and genetic variation in amphibian populations of the southeastern United States

North American amphibians have recently been impacted by two major emerging pathogens, the fungus Batrachochytrium dendrobatidis (Bd) and iridoviruses in the genus Ranavirus (Rv). Environmental factors and host genetics may play important roles in disease dynamics, but few studies incorporate both of these components into their analyses.

Here, we investigated the role of environmental and genetic factors in driving Bd and Rv infection prevalence and severity in a biodiversity hot spot, the southeastern United States.

We used quantitative PCR to characterize Bd and Rv dynamics in natural populations of three amphibian species: Notophthalmus perstriatus, Hyla squirella and Pseudacris ornata. We combined pathogen data, genetic diversity metrics generated from neutral markers, and environmental variables into general linear models to evaluate how these factors impact infectious disease dynamics. Occurrence, prevalence and intensity of Bd and Rv varied across species and populations, but only one species, Pseudacris ornata, harbored high Bd intensities in the majority of sampled populations. Genetic diversity and climate variables both predicted Bd prevalence, whereas climatic variables alone predicted infection intensity.

We conclude that Bd is more abundant in the southeastern United States than previously thought and that genetic and environmental factors are both important for predicting amphibian pathogen dynamics. Incorporating both genetic and environmental information into conservation plans for amphibians is necessary for the development of more effective management strategies to mitigate the impact of emerging infectious diseases.

Nuclear and Mitochondrial DNA Analyses of Golden Eagles (Aquila chrysaetos canadensis) from Three Areas in Western North America; Initial Results and Conservation Implications

Understanding the genetics of a population is a critical component of developing conservation strategies. We used archived tissue samples from golden eagles (Aquila chrysaetos canadensis) in three geographic regions of western North America to conduct a preliminary study of the genetics of the North American subspecies, and to provide data for United States Fish and Wildlife Service (USFWS) decision-making for golden eagle management. We used a combination of mitochondrial DNA (mtDNA) D-loop sequences and 16 nuclear DNA (nDNA) microsatellite loci to investigate the extent of gene flow among our sampling areas in Idaho, California and Alaska and to determine if we could distinguish birds from the different geographic regions based on their genetic profiles. Our results indicate high genetic diversity, low genetic structure and high connectivity. Nuclear DNA Fst values between Idaho and California were low but significantly different from zero (0.026). Bayesian clustering methods indicated a single population, and we were unable to distinguish summer breeding residents from different regions. Results of the mtDNA AMOVA showed that most of the haplotype variation (97%) was within the geographic populations while 3% variation was partitioned among them. One haplotype was common to all three areas. One region-specific haplotype was detected in California and one in Idaho, but additional sampling is required to determine if these haplotypes are unique to those geographic areas or a sampling artifact. We discuss potential sources of the high gene flow for this species including natal and breeding dispersal, floaters, and changes in migratory behavior as a result of environmental factors such as climate change and habitat alteration. Our preliminary findings can help inform the USFWS in development of golden eagle management strategies and provide a basis for additional research into the complex dynamics of the North American subspecies.

Quantifying the spatiotemporal dynamics in a chorus frog (Pseudacris) hybrid zone over 30 years

Although theory suggests that hybrid zones can move or change structure over time, studies supported by direct empirical evidence for these changes are relatively limited. We present a spatiotemporal genetic study of a hybrid zone between Pseudacris nigrita and P. fouquettei across the Pearl River between Louisiana and Mississippi. This hybrid zone was initially characterized in 1980 as a narrow and steep “tension zone,” in which hybrid populations were inferior to parentals and were maintained through a balance between selection and dispersal. We reanalyzed historical tissue samples and compared them to samples of recently collected individuals using microsatellites. Clinal analyses indicate that the cline has not shifted in roughly 30 years but has widened significantly. Anthropogenic and natural changes may have affected selective pressure or dispersal, and our results suggest that the zone may no longer best be described as a tension zone. To the best of our knowledge, this study provides the first evidence of significant widening of a hybrid cline but stasis of its center. Continued empirical study of dynamic hybrid zones will provide insight into the forces shaping their structure and the evolutionary potential they possess for the elimination or generation of species.

Permeability of the landscape matrix between amphibian breeding sites

For organisms that reproduce in discrete habitat patches, land cover between patches (known as the matrix) is important for dispersal among breeding sites. Models of patchy populations often incorporate information on the permeability of the matrix to dispersal, sometimes based on expert opinion. I estimated the relative resistance to gene flow of land cover types and barriers using F(ST) calculated from microsatellite markers in two amphibians, within an 800-km(2) area in northern Switzerland. The species included a frog (Rana temporaria: 996 individuals, 48 populations, seven markers) and a newt (Triturus alpestris: 816 individuals, 41 populations, seven markers). Open fields and urban areas were more resistant to gene flow than forested land; roads and highways also reduced permeability. Results were similar for the two species. However, differences in resistance among matrix elements were relatively low: gene flow through urban areas was reduced by only 24-42% relative to forest; a divided highway reduced gene flow by 11-40% and was 7-8 times more resistant than a secondary road. These data offer an empirically based alternative to expert opinion for setting relative resistance values in landscape models.