Multilocus estimation of genetic structure within populations

Stronger genetic differentiation among within-population genetic groups than among populations in Scots pine provides new insights into within-population genetic structuring

We investigated the presence of spatial genetic groups within forest tree populations and determined if the genetic divergence among these groups is greater than that between populations using Scots pine (Pinus sylvestris) as a model species. We genotyped 890 adult trees of Scots pine in six natural populations in Lithuania at 11 nuclear microsatellite loci. We used a Bayesian clustering approach to identify the within-population genetic groups within each of the six populations. We calculated the differentiation indexes among the genetic groups within each population and among the six populations by ignoring the genetic groups. The Bayesian clustering revealed 2 to 6 distinct genetic groups of varying size as the most likely genetic structures within populations. The genetic differentiation indexes among the genetic groups within populations were nearly tenfold greater (FST = 0.012-0.070) than those between the populations (FST = 0.003). We conclude on the existence of markedly stronger structuring of genetic variation within populations than between populations of Scots pine in large forest tracts of northern Europe. Such genetic structures serve as a contributing factor to large within population genetic diversity in northern conifers. We assume that within population mating in Scots pine is not completely random but rather is stratified into genetic clusters. Our study provides pioneering novel key insights into structuring of genetic variation within populations. Our findings have implications for examining within-population genetic diversity and genetic structure, conservation, and management of genetic resources.

Genetic Diversity and Population Structure of Dromedary Camel-Types

The dromedary camel is a unique livestock for its adaptations to arid-hot environments and its ability to provide goods under extreme conditions. There are no registries or breed standards for camels. Thus, named camel populations (i.e., camel-types) were examined for genetic uniqueness and breed status. Camel populations are generally named based on shared phenotype, country or region of origin, tribal ownership, or the ecology of their habitat. A dataset of 10 Short-Tandem Repeat markers genotyped for 701 individual camels from 27 camel-types was used to quantify genetic diversity within camel-types, compare genetic diversity across camel-types, determine the population genetic structure of camel-types, and identify camel-types that may represent true breeds. Summary statistics (genotyping call rate, heterozygosity, inbreeding coefficient FIS, and allelic frequencies) were calculated and population-specific analyses (pairwise FST, neighbor-joining tree, relatedness, Nei's genetic distance, principal coordinate analysis [PCoA], and STRUCTURE) were performed. The most notable findings were 1) little variation in genetic diversity was found across the camel-types, 2) the highest genetic diversity measure was detected in Targui and the lowest was in Awarik, 3) camel-types from Asia (especially the Arabian Peninsula) exhibited higher genetic diversity than their counterparts in Africa, 4) the highest DeltaK value of population structure separated camel-types based on geography (Asia vs. Africa), 5) the most distinct camel-types were the Omani, Awarik, and the Gabbra, 6) camel-types originating from the same country did not necessarily share high genetic similarity (e.g., camel-types from Oman), and 7) camel-type names were not consistently indicative of breed status.

Fine-scale analysis reveals cryptic landscape genetic structure in desert tortoises

Characterizing the effects of landscape features on genetic variation is essential for understanding how landscapes shape patterns of gene flow and spatial genetic structure of populations. Most landscape genetics studies have focused on patterns of gene flow at a regional scale. However, the genetic structure of populations at a local scale may be influenced by a unique suite of landscape variables that have little bearing on connectivity patterns observed at broader spatial scales. We investigated fine-scale spatial patterns of genetic variation and gene flow in relation to features of the landscape in desert tortoise (Gopherus agassizii), using 859 tortoises genotyped at 16 microsatellite loci with associated data on geographic location, sex, elevation, slope, and soil type, and spatial relationship to putative barriers (power lines, roads). We used spatially explicit and non-explicit Bayesian clustering algorithms to partition the sample into discrete clusters, and characterize the relationships between genetic distance and ecological variables to identify factors with the greatest influence on gene flow at a local scale. Desert tortoises exhibit weak genetic structure at a local scale, and we identified two subpopulations across the study area. Although genetic differentiation between the subpopulations was low, our landscape genetic analysis identified both natural (slope) and anthropogenic (roads) landscape variables that have significantly influenced gene flow within this local population. We show that desert tortoise movements at a local scale are influenced by features of the landscape, and that these features are different than those that influence gene flow at larger scales. Our findings are important for desert tortoise conservation and management, particularly in light of recent translocation efforts in the region. More generally, our results indicate that recent landscape changes can affect gene flow at a local scale and that their effects can be detected almost immediately.

Genetic structure and diversity of wild sorghum populations (Sorghum spp.) from different eco-geographical regions of Kenya

Wild sorghums are extremely diverse phenotypically, genetically and geographically. However, there is an apparent lack of knowledge on the genetic structure and diversity of wild sorghum populations within and between various eco-geographical regions. This is a major obstacle to both their effective conservation and potential use in breeding programs. The objective of this study was to assess the genetic diversity and structure of wild sorghum populations across a range of eco-geographical conditions in Kenya. Sixty-two wild sorghum populations collected from the 4 main sorghum growing regions in Kenya were genotyped using 18 simple sequence repeat markers. The study showed that wild sorghum is highly variable with the Coast region displaying the highest diversity. Analysis of molecular variance showed a significant variance component within and among wild sorghum populations within regions. The genetic structure of wild sorghum populations indicated that gene flow is not restricted to populations within the same geographic region. A weak regional differentiation was found among populations, reflecting human intervention in shaping wild sorghum genetic structure through seed-mediated gene flow. The sympatric occurrence of wild and cultivated sorghums coupled with extensive seed-mediated gene flow, suggests a potential crop-to-wild gene flow and vice versa across the regions. Wild sorghum displayed a mixed mating system. The wide range of estimated outcrossing rates indicate that some environmental conditions may exist where self-fertilisation is favoured while others cross-pollination is more advantageous.

Genetic structure and relationships within and between cultivated and wild sorghum (Sorghum bicolor (L.) Moench) in Kenya as revealed by microsatellite markers

Understanding the extent and partitioning of diversity within and among crop landraces and their wild/weedy relatives constitutes the first step in conserving and unlocking their genetic potential. This study aimed to characterize the genetic structure and relationships within and between cultivated and wild sorghum at country scale in Kenya, and to elucidate some of the underlying evolutionary mechanisms. We analyzed at total of 439 individuals comprising 329 cultivated and 110 wild sorghums using 24 microsatellite markers. We observed a total of 295 alleles across all loci and individuals, with 257 different alleles being detected in the cultivated sorghum gene pool and 238 alleles in the wild sorghum gene pool. We found that the wild sorghum gene pool harbored significantly more genetic diversity than its domesticated counterpart, a reflection that domestication of sorghum was accompanied by a genetic bottleneck. Overall, our study found close genetic proximity between cultivated sorghum and its wild progenitor, with the extent of crop-wild divergence varying among cultivation regions. The observed genetic proximity may have arisen primarily due to historical and/or contemporary gene flow between the two congeners, with differences in farmers' practices explaining inter-regional gene flow differences. This suggests that deployment of transgenic sorghum in Kenya may lead to escape of transgenes into wild-weedy sorghum relatives. In both cultivated and wild sorghum, genetic diversity was found to be structured more along geographical level than agro-climatic level. This indicated that gene flow and genetic drift contributed to shaping the contemporary genetic structure in the two congeners. Spatial autocorrelation analysis revealed a strong spatial genetic structure in both cultivated and wild sorghums at the country scale, which could be explained by medium- to long-distance seed movement.

Spatial correlations at different spatial scales are themselves highly correlated in isolation by distance processes

Although many properties of spatial autocorrelation statistics are well characterized, virtually nothing is known about possible correlations among values at different spatial scales, which ultimately would influence how inferences about spatial genetics are made at multiple spatial scales. This article reports the results of stochastic space-time simulations of isolation by distance processes, having a very wide range of amounts of dispersal for plants or animals, and analyses of the correlations among Moran's I-statistics for different mutually exclusive distance classes. In general, the stochastic correlations are extremely large (>0.90); however, the correlations bear a complex relationship with level of dispersal, spatial scale and spatial lag between distance classes. The correlations are so large that any existing or conceived statistical method that employs more than one distance class (or spatial scale) should not ignore them. This result also suggests that gains in statistical power via increasing sample size are limited, and that increasing numbers of assayed loci generally should be preferred. To the extent that sampling error for real data sets can be treated as white noise, it should be possible to account for stochastic correlations in formulating more precise statistical methods. Further, while the current results are for isolation by distance processes, they provide some guidance for some more complex stochastic space-time processes of landscape genetics. Moreover, the results hold for several popular measures other than Moran's I. In addition, in the results, the signal to noise ratios strongly decreased with distance, which also has several implications for optimal statistical methods using correlations at multiple spatial scales.

Effects of historical demography and ecological context on spatial patterns of genetic diversity within foxtail pine (Pinus balfouriana; Pinaceae) stands located in the Klamath Mountains, California

The density and dispersion of individuals, nonequilibrium demographics, and habitat fragmentation all affect the magnitude and extent of spatial genetic structure within forest tree populations. Here, we investigate the link between historical demography and spatial genetic structure within ecologically contrasting stands of foxtail pine (Pinus balfouriana) in the Klamath Mountains of northern California. We defined two stand types a priori, based largely on differences in foxtail pine density and basal area, and for each type we sampled two stands. Population expansions, likely from Pleistocene bottlenecks, were detected in three of the four stands. The magnitude and extent of spatial autocorrelation among genotypes at five nuclear microsatellites differed dramatically among stands, with those having lower foxtail pine density exhibiting strong patterns of isolation by distance. Moran's I statistics were 7-fold higher for the first distance class (<25 m) in these stands relative to those observed in stands with higher foxtail pine density (I(25) = 0.14 vs. 0.02). We conclude that differences in spatial genetic structure between stand types are due to differences in ecological attributes that affected expansion from inferred bottlenecks.

The impact of landscape disturbance on spatial genetic structure in the Guanacaste tree, Enterolobium cyclocarpum (Fabaceae)

We examined spatial genetic structure (SGS) in Enterolobium cyclocarpum (the Guanacaste tree), a dominant tree of Central American dry forests in 4 sites in Guanacaste Province, Costa Rica. In disturbed dry forest sites (e.g., pastures), E. cyclocarpum is primarily dispersed by cattle and horses, whose movements are restricted by pasture boundaries. The study sites varied in tree densities and disturbance. Allozyme analyses of adult trees demonstrated significant levels of SGS in 3 of 4 sites. SGS was primarily due to clusters of young adults located along seasonal streams, rocky areas, and in abandoned pastures. SGS was highest in the first distance class in the least disturbed population, which also had the lowest density of large adults. Low, but significant SGS characterized the site with the highest number of large adults located in individual pastures. The semiurban site, had no clusters of young adults and, probably as a result, failed to exhibit SGS. Our results demonstrate that disturbance can strongly influence SGS patterns and are consistent with a landscape model in which the location of potential recruitment sites, restricted seed disperser movements, and the number and location of maternal individuals dictate the level and pattern of SGS.

A heterogeneity test for fine-scale genetic structure

For organisms with limited vagility and/or occupying patchy habitats, we often encounter nonrandom patterns of genetic affinity over relatively small spatial scales, labelled fine-scale genetic structure. Both the extent and decay rate of that pattern can be expected to depend on numerous interesting demographic, ecological, historical, and mating system factors, and it would be useful to be able to compare different situations. There is, however, no heterogeneity test currently available for fine-scale genetic structure that would provide us with any guidance on whether the differences we encounter are statistically credible. Here, we develop a general nonparametric heterogeneity test, elaborating on standard autocorrelation methods for pairs of individuals. We first develop a 'pooled within-population' correlogram, where the distance classes (lags) can be defined as functions of distance. Using that pooled correlogram as our null-hypothesis reference frame, we then develop a heterogeneity test of the autocorrelations among different populations, lag-by-lag. From these single-lag tests, we construct an analogous test of heterogeneity for multilag correlograms. We illustrate with a pair of biological examples, one involving the Australian bush rat, the other involving toadshade trillium. The Australian bush rat has limited vagility, and sometimes occupies patchy habitat. We show that the autocorrelation pattern diverges somewhat between continuous and patchy habitat types. For toadshade trillium, clonal replication in Piedmont populations substantially increases autocorrelation for short lags, but clonal replication is less pronounced in mountain populations. Removal of clonal replicates reduces the autocorrelation for short lags and reverses the sign of the difference between mountain and Piedmont correlograms.

Comparative analysis of the within-population genetic structure in wild cherry (Prunus avium L.) at the self-incompatibility locus and nuclear microsatellites

Gametophytic self-incompatibility (SI) systems in plants exhibit high polymorphism at the SI controlling S-locus because individuals with rare alleles have a higher probability to successfully pollinate other plants than individuals with more frequent alleles. This process, referred to as frequency-dependent selection, is expected to shape number, frequency distribution, and spatial distribution of self-incompatibility alleles in natural populations. We investigated the genetic diversity and the spatial genetic structure within a Prunus avium population at two contrasting gene loci: nuclear microsatellites and the S-locus. The S-locus revealed a higher diversity (15 alleles) than the eight microsatellites (4-12 alleles). Although the frequency distribution of S-alleles differed significantly from the expected equal distribution, the S-locus showed a higher evenness than the microsatellites (Shannon's evenness index for the S-locus: E = 0.91; for the microsatellites: E = 0.48-0.83). Also, highly significant deviations from neutrality were found for the S-locus whereas only minor deviations were found for two of eight microsatellites. A comparison of the frequency distribution of S-alleles in three age-cohorts revealed no significant differences, suggesting that different levels of selection acting on the S-locus or on S-linked sites might also affect the distribution and dynamics of S-alleles. Autocorrelation analysis revealed a weak but significant spatial genetic structure for the multilocus average of the microsatellites and for the S-locus, but could not ascertain differences in the extent of spatial genetic structure between these locus types. An indirect estimate of gene dispersal, which was obtained to explain this spatial genetic pattern, indicated high levels of gene dispersal within our population (sigma(g) = 106 m). This high gene dispersal, which may be partly due to the self-incompatibility system itself, aids the effective gene flow of the microsatellites, thereby decreasing the contrast between the neutral microsatellites and the S-locus.

Microsatellites for ecologists: a practical guide to using and evaluating microsatellite markers

Recent improvements in genetic analysis and genotyping methods have resulted in a rapid expansion of the power of molecular markers to address ecological questions. Microsatellites have emerged as the most popular and versatile marker type for ecological applications. The rise of commercial services that can isolate microsatellites for new study species and genotype samples at reasonable prices presents ecologists with the unprecedented ability to employ genetic approaches without heavy investment in specialized equipment. Nevertheless, the lack of accessible, synthesized information on the practicalities and pitfalls of using genetic tools impedes ecologists' ability to make informed decisions on using molecular approaches and creates the risk that some will use microsatellites without understanding the steps needed to evaluate the quality of a genetic data set. The first goal of this synthesis is to provide an overview of the strengths and limitations of microsatellite markers and the risks, cost and time requirements of isolating and using microsatellites with the aid of commercial services. The second goal is to encourage the use and consistent reporting of thorough marker screening to ensure high quality data. To that end, we present a multistep screening process to evaluate candidate loci for inclusion in a genetic study that is broadly targeted to both novice and experienced geneticists alike.

Optimal sampling strategy for estimation of spatial genetic structure in tree populations

Fine-scale spatial genetic structure (SGS) in natural tree populations is largely a result of restricted pollen and seed dispersal. Understanding the link between limitations to dispersal in gene vectors and SGS is of key interest to biologists and the availability of highly variable molecular markers has facilitated fine-scale analysis of populations. However, estimation of SGS may depend strongly on the type of genetic marker and sampling strategy (of both loci and individuals). To explore sampling limits, we created a model population with simulated distributions of dominant and codominant alleles, resulting from natural regeneration with restricted gene flow. SGS estimates from subsamples (simulating collection and analysis with amplified fragment length polymorphism (AFLP) and microsatellite markers) were correlated with the 'real' estimate (from the full model population). For both marker types, sampling ranges were evident, with lower limits below which estimation was poorly correlated and upper limits above which sampling became inefficient. Lower limits (correlation of 0.9) were 100 individuals, 10 loci for microsatellites and 150 individuals, 100 loci for AFLPs. Upper limits were 200 individuals, five loci for microsatellites and 200 individuals, 100 loci for AFLPs. The limits indicated by simulation were compared with data sets from real species. Instances where sampling effort had been either insufficient or inefficient were identified. The model results should form practical boundaries for studies aiming to detect SGS. However, greater sample sizes will be required in cases where SGS is weaker than for our simulated population, for example, in species with effective pollen/seed dispersal mechanisms.

Estimating dispersal from short distance spatial autocorrelation

A series of theoretical studies has formed a strong connection between spatial statistics observed in populations and summary measures of the amount of dispersal. Synthesized, these developments allow dispersal to be indirectly estimated from standing spatial patterns of genetic variation under a range of conditions broad enough to be likely met in most populations of either plants or animals. The spatial correlations at the shortest distances are particularly robust to range of conditions and have disproportionately high statistical power. This review integrates theoretical results in a way that maximizes robustness and flexibility in the use of short distance autocorrelation to estimate Wright's neighborhood size, or the total variance in dispersal distances. Empirical guidelines are developed that are meant to be as practical and broad as possible. The guidelines focus on Moran's I-statistics for diploid genotypes converted to allele frequencies, but are also extended to or compared with several other approaches.

Spatial and population genetic structure of microsatellites in white pine

We evaluated the population genetic structure of seven microsatellite loci for old growth and second growth populations of eastern white pine (Pinus strobus). From each population, located within Hartwick Pines State Park, Grayling, Michigan, USA, 120-122 contiguous trees were sampled for genetic analysis. Within each population, genetic diversity was high and inbreeding low. When comparing these populations, there is a significant, but small (less than 1%), genetic divergence between populations. Spatial distance between populations or timber harvest at the second growth site were reasonable explanations for the observed minor differences in allele frequencies between populations. Spatial autocorrelation analysis suggested that, for the old growth population, weak positive structuring at 15 m fits the isolation by distance model for a neighbourhood size of about 100 individuals. In comparison, genotypes were randomly distributed in the second growth population. Thus, logging may have decreased spatial structuring at the second growth site, suggesting that management practices may be used to alter natural spatial patterns. In addition, the amount of autocorrelation in the old growth population appears to be lower for some of the microsatellites, suggesting higher numbers of rare alleles and that higher mutation rates may have directly affected spatial statistics by reducing structure.

Mutation at high rates reduces spatial structure within populations

The question of whether or not the high rates (mu) of mutation that occur for some hypervariable markers can affect commonly used empirical measures of spatial structure of genetic variation within populations is addressed. The results show that values of these measures are approximately halved when mu is 10(-2). Finest spatial-scale correlations, measured by either Moran's I-statistics or conditional kinship, are reduced by 30%-50%. When the mutation rate is 10 times lower, much smaller reductions result, e.g. averaging 7% for the finest scale correlations. Still smaller orders of magnitude of mu cause negligible changes in spatial structure, where any effects normally would not be detectable. The reductions are caused by forward mutations, and when the reductions are measured as percentages, they are nearly independent of the amount of structure produced sans mutation, except when dispersal is nearly minimal. The percent reductions are also nearly independent of the number of alleles and of back mutations, hence of the nature of the mutation process (e.g. stepwise or not). The results demonstrate that some hypervariable loci should have reduced spatial structuring, and that marker choice may affect the values observed in experimental surveys. Moreover, if fine-scale correlations are used to indirectly estimate dispersal distances, then mutation at high rates could inflate estimates, easily up to two- to three-fold.