Basics of Spatial Data Analysis: Linking Landscape and Genetic Data for Landscape Genetic Studies

Landscape genetics is an important framework for investigating the influence of spatial pattern on ecological process. Nevertheless, the standard analytic frameworks in landscape genetics have difficulty evaluating hypotheses about spatial processes in dynamic landscapes. We use a predictive hypothesis-driven approach to quantify the relative contribution of historic and contemporary processes to genetic connectivity. By confronting genetic data with models of historic and contemporary landscapes, we identify dispersal processes operating in naturally heterogeneous and human-altered systems. We demonstrate the approach using a case study of microsatellite polymorphism and indirect estimates of gene flow for a rainforest bird, the logrunner (Orthonyx temminckii). Of particular interest was how much information in the genetic data was attributable to processes occurring in the reconstructed historic landscape and contemporary human-modified landscape. A linear mixed model was used to estimate appropriate sampling variance from nonindependent data and information-theoretic model selection provided strength of evidence for alternative hypotheses. The contemporary landscape explained slightly more information in the genetic differentiation data than the historic landscape, and there was considerable evidence for a temporal shift in dispersal pattern. In contrast, migration rates estimated from genealogical information were primarily influenced by contemporary landscape change. We discovered that landscape heterogeneity facilitated gene flow before European settlement, but contemporary deforestation is rapidly becoming the most important barrier to logrunner dispersal.

The integration of ecology and genetics has become established in recent decades, in hand with the development of new technologies, whose implementation is allowing an improvement of the tools used for data analysis. In a landscape genetics context, integrative management of population information from different sources can make spatial studies involving phenotypic, genotypic and environmental data simpler, more accessible and faster. Tools for exploratory analysis of autocorrelation can help to uncover the spatial genetic structure of populations and generate appropriate hypotheses in searching for possible causes and consequences of their spatial processes. This study presents EcoGenetics, an R package with tools for multisource management and exploratory analysis in landscape genetics.

Prediction and Prevention of Parasitic Diseases Using a Landscape Genomics Framework

In landscape genetics, isolation-by-distance (IBD) is regarded as a baseline pattern that is obtained without additional effects of landscape elements on gene flow. However, the configuration of suitable habitat patches determines deme topology, which in turn should affect rates of gene flow. IBD patterns can be characterized either by monotonically increasing pairwise genetic differentiation (for example, FST) with increasing interdeme geographic distance (case-I pattern) or by monotonically increasing pairwise genetic differentiation up to a certain geographical distance beyond which no correlation is detectable anymore (case-IV pattern). We investigated if landscape configuration influenced the rate at which a case-IV pattern changed to a case-I pattern. We also determined at what interdeme distance the highest correlation was measured between genetic differentiation and geographic distance and whether this distance corresponded to the maximum migration distance. We set up a population genetic simulation study and assessed the development of IBD patterns for several habitat configurations and maximum migration distances. We show that the rate and likelihood of the transition of case-IV to case-I FST-distance relationships was strongly influenced by habitat configuration and maximum migration distance. We also found that the maximum correlation between genetic differentiation and geographic distance was not related to the maximum migration distance and was measured across all deme pairs in a case-I pattern and, for a case-IV pattern, at the distance where the FST-distance curve flattens out. We argue that in landscape genetics, separate analyses should be performed to either assess IBD or the landscape effects on gene flow.

Wing geometry and genetic analyses reveal contrasting spatial structures between male and female Aedes aegypti (L.) (Diptera: Culicidae) populations in metropolitan Manila, Philippines.

This PhD Thesis includes five main parts on diverse topics. The first two parts deal with the trophic ecology of wolves in Italy consequently to a recent increase of wild ungulates abundance. Data on wolf diet across time highlighted how wild ungulates are important food resource for wolves in Italy. Increasing wolf population, increasing numbers of wild ungulates and decreasing livestock consume are mitigating wolf-man conflicts in Italy in the near future. In the third part, non-invasive genetic sampling techniques were used to obtain genotypes and genders of about 400 wolves. Thus, wolf packs were genetically reconstructed using diverse population genetic and parentage software. Combining the results on pack structure and genetic relatedness with sampling locations, home ranges of wolf packs and dispersal patterns were identified. These results, particularly important for the conservation management of wolves in Italy, illustrated detailed information that can be retrieved from genetic identification of individuals. In the fourth part, wolf locations were combined with environmental information obtained as GIS-layers. Modern species distribution models (niche models) were applied to infer potential wolf distribution and predation risk. From the resulting distribution maps, information pastures with the highest risk of depredation were derived. This is particularly relevant as it allows identifying those areas under danger of carnivore attack on livestock. Finally, in the fifth part, habitat suitability models were combined with landscape genetic analysis. On one side landscape genetic analyses on the Italian wolves provided new information on the dynamics and connectivity of the population and, on the other side, a profound analysis of the effects that habitat suitability methods had on the parameterization of landscape genetic analyses was carried out to contributed significantly to landscape genetic theory.

Characterizing genetic diversity within and between populations has broad applications in studies of human disease and evolution. We propose a new approach, spatial ancestry analysis, for the modeling of genotypes in two- or three-dimensional space. In spatial ancestry analysis (SPA), we explicitly model the spatial distribution of each SNP by assigning an allele frequency as a continuous function in geographic space. We show that the explicit modeling of the allele frequency allows individuals to be localized on the map on the basis of their genetic information alone. We apply our SPA method to a European and a worldwide population genetic variation data set and identify SNPs showing large gradients in allele frequency, and we suggest these as candidate regions under selection. These regions include SNPs in the well-characterized LCT region, as well as at loci including FOXP2, OCA2 and LRP1B.

Habitat loss, fragmentation and degradation are drivers of major declines in biodiversity and species extinctions. The actual causes of species population declines following habitat change are more difficult to discern. In this thesis I attempt to identify, based on published data on mobility and their ecological responses to fragmentation, some of the processes that may have led to patterns of decline for 10 woodland-dependent birds in fragmented woodlands of central Victoria, Australia. Eight species had been identified as ‘decliners’ (species that disappear from suitable patches when landscape-level tree cover falls below species-specific thresholds) and two as ‘tolerant’ species (whose occurrence in suitable habitat patches is independent of landscape tree cover). I investigated the contribution of decreased structural connectivity on functional connectivity as a cause of the observed declines. A set of landscape connectivity models was constructed of each species covering a range of plausible values. The two dominant algorithms for summarising effective distances used in modelling complex fragmentation patterns in landscape genetics—least-cost path and circuit distance—were used to construct these models The results of the two methods were compared and circuit distance was determined to be the more appropriate approach for use in my study system. Choice of algorithm and null model were important influences on inferences in landscape genetics. I predicted (1) fragmentation would impede dispersal and gene flow of ‘decliners’ but not of ‘tolerant’ species; and that fragmentation effects would be stronger (2) in the least mobile species, (3) in the more philopatric sex and (4) in the more fragmented region. These predictions were then tested with a large empirical genetic dataset (2198 individuals from 63 sites across a 170 x 50-km study area). I fitted models specific to sex and geographic zone in order to account for sex-biased dispersal and potential scale- and configuration-specific effects. As expected, four of the least mobile decliners showed reduced genetic connectivity. Responses were sex specific in the two least mobile species. The tolerant species and (unexpectedly) four of the more mobile decliners showed no reduction in gene flow. Weaker genetic effects were observed in the geographic zone with more aggregated vegetation, consistent with gene flow being unimpeded by landscape structure. These results indicate that, excepting the most sedentary species in our system, the movement of the more dispersive sex maintains overall genetic connectivity across fragmented landscapes in the study area. I examined relationships among configuration, extent and status of native vegetation and three commonly used indicators of individual body condition and chronic stress in 13 species, two measures of changes to population processes (sex ratio and individual homozygosity) in 10 species and allelic richness in five species. Little support for relationships between site or landscape characteristics and individual or population response variables was found. These findings, along with related work to which I contributed, but that does not form part of the thesis, highlighted the need for management to increase both connectivity, for the least mobile species, and critical resource availability for other species to conserve these declining species.

One of the pervasive challenges in landscape genetics is detecting gene flow patterns within continuous populations of highly mobile wildlife. Understanding population genetic structure within a continuous population can give insights into social structure, movement across the landscape and contact between populations, which influence ecological interactions, reproductive dynamics or pathogen transmission. We investigated the genetic structure of a large population of deer spanning the area of Wisconsin and Illinois, USA, affected by chronic wasting disease. We combined multiscale investigation, landscape genetic techniques and spatial statistical modelling to address the complex questions of landscape factors influencing population structure. We sampled over 2000 deer and used spatial autocorrelation and a spatial principal components analysis to describe the population genetic structure. We evaluated landscape effects on this pattern using a spatial autoregressive model within a model selection framework to test alternative hypotheses about gene flow. We found high levels of genetic connectivity, with gradients of variation across the large continuous population of white-tailed deer. At the fine scale, spatial clustering of related animals was correlated with the amount and arrangement of forested habitat. At the broader scale, impediments to dispersal were important to shaping genetic connectivity within the population. We found significant barrier effects of individual state and interstate highways and rivers. Our results offer an important understanding of deer biology and movement that will help inform the management of this species in an area where overabundance and disease spread are primary concerns.

Maintaining tiger connectivity and minimizing extinction into the next century: Insights from landscape genetics and spatially-explicit simulations

The processes by which plant spatial patterns are formed, and the effects of those patterns on plant community dynamics, remain important areas of research in plant ecology. Plant spatial pattern formation has been linked to many ecological processes that act to structure plant communities at different spatiotemporal scales. Past studies of pattern formation are common, but recent methodological advances in data collection and analysis have permitted researchers to conduct more advanced observational studies of pattern formation in space and time. While studies of the effects of plant spatial patterns were formally rare, they have increased in the last decade as new types of experiments and analysis have been developed to better understand the myriad effects of plant patterns on community dynamics. My dissertation research examined both the causes and consequences of plant spatial patterns in the context of natural and experimental Great Basin semi-arid plant communities. In both cases, I implemented novel methodologies for data collection, experimental design, and data analysis in an attempt to address current gaps in knowledge related to the processes by which plant spatial patterns are formed, as well as the effect of plant spatial patterns on community dynamics. The results inform both basic and applied plant ecology, and set the stage for further research on the causes and consequences of plant spatial patterns in semi-arid plant communities.

The available scientific literature was reviewed to assess the taxonomic standing of North American wolves, including subspecies of the gray wolf, Canis lupus. The recent scientific proposal that the eastern wolf, C. l. lycaon, is not a subspecies of gray wolf, but a full species, Canis lycaon, is well-supported by both morphological and genetic data. This species' range extends westward to Minnesota, and it hybridizes with gray wolves where the two species are in contact in eastern Canada and the Upper Peninsula of Michigan, Wisconsin, and Minnesota. Genetic data support a close relationship between eastern wolf and red wolf Canis rufus, but do not support the proposal that they are the same species; it is more likely that they evolved independently from different lineages of a common ancestor with coyotes. The genetic distinctiveness of the Mexican wolf Canis lupus baileyi supports its recognition as a subspecies. The available genetic and morphometric data do not provide clear support for the recognition of the Arctic wolf Canis lupus arctos, but the available genetic data are almost entirely limited to one group of genetic markers (microsatellite DNA) and are not definitive on this question. Recognition of the northern timber wolf Canis lupus occidentalis and the plains wolf Canis lupus nubilus as subspecies is supported by morphological data and extensive studies of microsatellite DNA variation where both subspecies are in contact in Canada. The wolves of coastal areas in southeastern Alaska and British Columbia should be assigned to C. lupus nubilus. There is scientific support for the taxa recognized here, but delineation of exact geographic boundaries presents challenges. Rather than sharp boundaries between taxa, boundaries should generally be thought of as intergrade zones of variable width.

Landscape genetics was developed to detect landscape elements shaping genetic population structure, including the effects of fragmentation. Multifarious environmental variables can influence gene flow in different ways and expert knowledge is frequently used to construct friction maps. However, the extent of the migration and the movement of single individuals are frequently unknown, especially for non-model species, and friction maps only based on expert knowledge can be misleading. In this study, we used three different methods: isolation by distance (IBD), least-cost modelling and a strip-based approach to disentangle the human implication in the fragmentation process in the slow worm (Anguis fragilis), as well as the specific landscape elements shaping the genetic structure in a highly anthropized 16 km2 area in Switzerland. Friction maps were constructed using expert opinion, but also based on the combination of all possible weightings for all landscape elements. The IBD indicated a significant effect of geographic distance on genetic differentiation. Further approaches demonstrated that highways and railways were the most important elements impeding the gene flow in this area. Surprisingly, we also found that agricultural areas and dense forests seemed to be used as dispersal corridors. These results confirmed that the slow worm has relatively unspecific habitat requirements. Finally, we showed that our models based on expert knowledge performed poorly compared to cautious analysis of each variable. This study demonstrated that landscape genetic analyses should take expert knowledge with caution and exhaustive analyses of each landscape element without a priori knowledge and different methods can be recommended.

As the field of landscape genetics is progressing toward comparative empirical studies and meta-analysis, it is important to know how best to compare the strength of spatial genetic structure between studies and species. Moran’s Eigenvector Maps are a promising method that does not make an assumption of isolation-by-distance in a homogeneous environment but can discern cryptic structure that may result from multiple processes operating in heterogeneous landscapes. MEMgene uses spatial filters from Moran’s Eigenvector Maps as predictor variables to explain variation in a genetic distance matrix, and it returns adjusted R2 as a measure of the amount of genetic variation that is spatially structured. However, it is unclear whether, and under which conditions, this value can be used to compare the degree of spatial genetic structure (effect size) between studies. This study addresses the fundamental question of comparability at two levels: between independent studies (meta-analysis mode) and between species sampled at the same locations (comparative mode). We used published datasets containing 9,900 haploid, biallelic, neutral loci simulated on a quasi-continuous, square landscape under four demographic scenarios (island model, isolation-by-distance, expansion from one or two refugia). We varied the genetic resolution (number of individuals and loci) and the number of random sampling locations. We considered two measures of effect size, the MEMgene adjusted R2 and multivariate Moran’s I, which is related to Moran’s Eigenvector Maps. Both metrics were highly sensitive to the number of locations, even when using standardized effect sizes, SES, and the number of individuals sampled per location, but not to the number of loci. In comparative mode, using the same Moran Eigenvector Maps for all species, even those with missing values at some sampling locations, reduced bias due to the number of locations under isolation-by-distance (stationary process) but increased it under expansion from one or two refugia (non-stationary process). More robust measures of effect size need to be developed before the strength of spatial genetic structure can be accurately compared, either in a meta-analysis of independent empirical studies or within a comparative, multispecies landscape genetic study.

Habitat destruction and fragmentation are major factors in the destruction of genetic diversity and affect the movement behavior of the Roe deer population in the remaining habitats. Here, we study the population and landscape genetics of Capreolus capreolus (roe deer) in northern and northwestern Iran using twelve polymorphism microsatellite markers. From 111 total specimens, 63 had successful extraction (6 feces, 35 tissues, 9 bones, and 13 antlers). We considered 30 microsatellite polymorphic loci, of which only 12 were amplified for our further analysis. For genetic diversity analysis, the Weir-Cockerham method was applied to measure the inbreeding coefficient (FIS) and fixation index (FST) for each locus as well as for each population. For landscape genetics, the susceptibility patterns of genetic variations were assessed using three hypotheses including isolation by distance (IBD), isolation by environment (IBE), isolation by resistance (IBR), and individual landscape genetic analysis. A habitat suitability map as an indicator of landscape resistance was constructed from several species distribution models (SDMs) algorithms including Generalized Boosting Models (GBM), Maximum Entropy (Maxent), Random Forest (RF), Generalized Linear Model (GLM), Multivariate Adaptive Regression Splines (MARS) and artificial neural networks (ANN) and an ensemble model. Our estimated FIs index showed that the Golestan, Arasbaran, and Guilan populations had the highest and lowest genetic diversity among roe deer populations. According to the Fst criterion, our results showed that Golestan and East Azarbaijan (Arasbaran) had the highest and Mazandaran had the lowest genetic distance patterns. Our results do not suggest that there is high genetic differentiation for roe deer in the region, with high levels of gene flow between study areas. We found that geographic distance has no significant relationship with genetic distance and that there is no significant relationship between the ecological niche non-similarity matrix and the genetic distance matrix. The most influential factors affecting gene flow in roe deer were aspect and elevation variables. The analysis suggests that the landscape has no significant influence on the structuring of the studied population and shows little genetic differentiation.