Microsatellite variation and differentiation among local populations of Castanopsis species in Japan

Microsatellite variations in Castanopsis species in Japan were examined to clarify the genetic relationships among 25 local populations according to the difference in the number of layers of adaxial epidermis in the leaves. Six microsatellite loci were assayed for 629 seedlings from the populations, and these seedlings were classified into five types according to the state of the leaf epidermis. Remarkable differences in the allele frequency of the six microsatellite loci were observed among these local populations. The coefficients of genetic differentiation, R_ST, of each locus ranged from 0.209 to 0.388. An unweighted pair-group method (UPGMA) phenogram constructed on the population pairwise R_ST over the loci revealed three clusters (A–C), and six sub-clusters. These clusters reflected the differences in the occurrence frequency of seedlings in each epidermis type within a population. Our findings suggest that clusters A and C are the local populations dominated by Castanopsis sieboldii and Castanopsis cuspidata, respectively, while local populations of cluster B are composed of the two Castanopsis species and/or include many individuals derived by hybridization. The six sub-clusters were found to reflect the geographic relationship among the populations, suggesting a different process for geographic population dynamics during the postglacial period.     

Low levels of quantitative and molecular genetic differentiation among natural populations of Medicago ciliaris Kroch. (Fabaceae) of different Tunisian eco-geographical origin

In this paper, we analyze the genetic variability in four Tunisian natural populations of Medicago ciliaris using 19 quantitative traits and six polymorphic microsatellite loci. We investigated the amplification transferability of 30 microsatellites developed in the model legume M. truncatula to M. ciliaris. Results revealed that about 56.66% of analyzed markers are valuable genetic markers for M. ciliaris. The most genetic diversity at quantitative traits and microsatellite loci was found to occur within populations (>80%). Low differentiations among populations at quantitative traits Q _ST  = 0.146 and molecular markers F _ST  = 0.18 were found. The majority of measured traits exhibited no significant difference in the level of Q _ST and F _ST . Furthermore, significant correlations established between these traits and eco-geographical factors suggested that natural selection should be invoked to explain the level of phenotypic divergence among populations rather than drift. There was no significant correlation between population differentiation at quantitative traits and molecular markers. Significant spatial genetic structure consistent with models of isolation by distance was detected within all studied populations. The site-of-origin environmental factors explain about 9.07% of total phenotypic genetic variation among populations. The eco-geographical factors that influence more the variation of measured traits among populations are the soil texture and altitude. Nevertheless, there were no consistent pattern of associations between gene diversity (He) and environmental factors.     

Morphological and microsatellite diversity associated with ecological factors in natural populations of Medicago laciniata Mill. (Fabaceae)

Genetic variability in 10 natural Tunisian populations of Medicago laciniata were analysed using 19 quantitative traits and 12 polymorphic microsatellite loci. A large degree of genetic variability within-populations and among-populations was detected for both quantitative characters and molecular markers. High genetic differentiation among populations for quantitative traits was seen, with Q _ST = 0.47, and F _ST = 0.47 for microsatellite markers. Several quantitative traits displayed no statistical difference in the levels of Q _ST and F _ST . Further, significant correlations between quantitative traits and eco-geographical factors suggest that divergence in the traits among populations may track environmental differences. There was no significant correlation between genetic variability at quantitative traits and microsatellite markers within populations. The site-of-origin of eco-geographical factors explain between 18.13% and 23.40% of genetic variance among populations at quantitative traits and microsatellite markers, respectively. The environmental factors that most influence variation in measured traits among populations are assimilated phosphorus (P₂0₅) and mean annual rainfall, followed by climate and soil texture, altitude and organic matter. Significant associations between eco-geographical factors and gene diversity, H _e , were established in five-microsatellite loci suggesting that these simple sequence repeats (SSRs) are not necessarily biologically neutral.     

Genetic structure of refugial populations of the temperate plant Shortia rotundifolia (Diapensiaceae) on a subtropical island

Continental island systems harbour relict biota and populations that might have migrated during glacial periods due to the formation of landbridges. Here we analysed the genetic structure of relict populations of the temperate plant Shortia rotundifolia on the subtropical island of Iriomotejima, Japan. This plant, which inhabits riparian environments, is designated “near threatened”. Only five extant populations have been found on the island. Our analyses of 10 nuclear microsatellite loci detected genetic diversity of H _E = 0.488 and H _O = 0.358 for all populations of S. rotundifolia on the island. A high inbreeding coefficient for all populations together (F _IS = 0.316) and each population separately (F _IS = 0.258–0.497) might be attributable to crossing among closely related descendants within a population, an idea that is supported by the relatedness coefficient. These results and an examination of the populations’ demographic histories suggest that the extant populations on Iriomotejima have not experienced a recent population bottleneck. The five extant populations were genetically differentiated (F _ST = 0.283; P < 0.001), suggesting low seed dispersal by gravity and/or low pollen flow via pollinators in the riparian environment. In addition, population differentiation was not related to genetic distance, implying that at one time, ancestral populations might have been distributed over a wider area of the island. However, population fragmentation and range contraction might have occurred at random during the postglacial period.     

Genetic structure of the Eurasian lynx population in north-eastern Poland and the Baltic states

We analyzed the genotypes of Eurasian lynx (Lynx lynx) from three populations in the westernmost part of the species main range. One population was situated at the distribution edge (NE Poland) and the two other (Latvia and Estonia) were located within the main, contiguous range of the species. The aim was to determine if the genetic composition varied among these populations and if there was evidence of isolation among them. Based on microsatellite allele frequencies, we found the allelic richness in Polish lynx to be lower than that in lynx from Latvia and Estonia. We also found significant differentiation among the lynx populations, with the NE Poland population forming a distinct genetic group relative to the two other populations (R _ST = 0.15 and 0.22, P < 0.0001). We suggest that genetic differentiation among lynx populations is the result of habitat insularisation that limits gene flow. This finding emphasizes the necessity to consider the lynx genetic differentiation in conservation planning of this species in Poland.     

Genetic divergence in nuclear genomes between populations of <Emphasis Type="Italic">Fagus crenata</Emphasis> along the Japan Sea and Pacific sides of Japan

Genetic diversity and structure in Fagus crenata were studied by analyzing 14 nuclear microsatellite loci in 23 populations distributed throughout the species’ range. Although population differentiation was very low (F _ST = 0.027; R _ST = 0.041), both neighbor-joining tree and Bayesian clustering analyses provided clear evidence of genetic divergence between populations along the Japan Sea (Japan Sea lineage) and Pacific (Pacific lineage) sides of Japan, indicating that physical barriers to migration and gene flow, notably the mountain ranges separating the populations along the Japan Sea and Pacific sides, have promoted genetic divergence between these populations. The two lineages of the nuclear genome are generally consistent with those of the chloroplast genome detected in a previous study, with several discrepancies between the two genomes. Within-population genetic diversity was generally very high (average H _E = 0.839), but decreased in a clinal fashion from southwest to northeast, largely among populations of the Japan Sea lineage. This geographical gradient may have resulted from the late-glacial and postglacial recolonization to the northeast, which led to a loss of within-population genetic diversity due to cumulative founder effects.     

Genetic diversity of the black mangrove (Avicennia germinans L.) in Colombia

This study analyzed the genetic diversity and patterns of genetic structure in Colombian populations of Avicennia germinans L. using microsatellite loci. A lower genetic diversity was found on both the Caribbean (Ho = 0.439) and the Pacific coasts (Ho = 0.277) than reported for the same species in other locations of Central American Pacific, suggesting the deterioration of genetic diversity. All the populations showed high inbreeding coefficients (0.131–0.462) indicating heterozygotes deficience. The genetic structure between the Colombian coasts separated by Central American Isthmus was high (F_RT = 0.39) and the analyses of the genetic patterns of A. germinans revealed a clear differentiation of populations and no-recent gene flow evidence between coasts. Genetic structure was found within each coast (F_ST = 0.10 for the Caribbean coast and F_ST = 0.22 for the Pacific coast). The genetic patterns along the two coasts appear to reflect a forcing by local geomorphology and marine currents. Both coasts constitute a different Evolutionary Significant Unit, so we suggest for future transplantations plans that propagules or saplings of the populations of the Caribbean coast should not be mixed with those of the Pacific Colombian coast. Besides, we suggest that reforestation efforts should carefully distinguish propagules sources within each coast.     

Genetic diversity and structure of the endemic Caesalpinia hintonii complex (Leguminosae: Caesalpinioideae) in Mexico

The Caesalpinia hintonii complex is formed by five endemic species (C. hintonii, C. laxa, C. macvaughii, C. melanadenia and C. epifanioi) occurring in central Mexico. This species complex is under incipient genetic divergence as by-product of local adaptations in reproductive and morphological traits to different habitats. We estimate the genetic variation and structure of populations of this species complex to assess the extent of genetic differentiation among populations and related species along its geographic distribution. Estimations of genetic diversity and structure were done based on ten enzymes and 18 loci. Mean number of alleles per locus ranged from 1.5 to 1.9. Polymorphic loci ranged from 42.1 to 68.4. Observed (H_o: range 0.191–0.275) and expected (H_e: range 0.205–0.317) heterozygosities in this complex were higher compared with other endemic and legume species. Neis genetic diversity estimates showed that most genetic variation was found within (H_S = 0.325) rather than among populations (D_ST = 0.085). Populations of the species C. hintonii showed a considerable genetic differentiation (F_ST = 0.207). The results of genetic diversity and structure within and among populations are in accord with the great morphological differentiation described for this species complex.     

Genetic variation amongst viviparid snails in the genus Mekongia (Prosobranchia: Viviparidae) in Thailand

Allozyme variation in viviparid snails from the genus Mekongia in Thailand were examined across the different species, subspecies and geographical locations (river drainage systems). Using horizontal starch gel electrophoresis, 11 presumed allozyme loci (eight polymorphic) from eight enzyme systems were screened. Heterozygosity was moderately low (Hexp = 0.000–0.109, mean = 0.037). One population of Mekongia pongensis (Nong Khai) was monomorphic at all 11 examined loci in contrast to the other two populations, suggesting bottleneck within this population. Populations were more differentiated in the Mekongia sphaericula complex (F_ST = 0.587) than in either the Mekongia swainsoni complex (F_ST = 0.161) or M. pongensis species (F_ST = 0.073). Mekongia sphaericula sphaericula and Mekongia sphaericula extensa exhibited fixed allele differences at two loci, a high genetic distance (D = 0.265–0.300) and a potential polyphyletic relationship, suggesting two distinct lineages (species).     

Phylogeography of two East Asian species in Croomia (Stemonaceae) inferred from chloroplast DNA and ISSR fingerprinting variation

The genus Croomia (Stemonaceae) comprises three herbaceous perennial species that are distributed in temperate-deciduous forests in Southeastern North America (C. pauciflora) and East Asia (C. japonica, C. heterosepala). The two Asian species have abutting ranges in South Japan, but C. japonica also occurs disjunctively on the adjacent Asiatic mainland in East China. In our phylogenetic analysis of Croomia, based on chloroplast (cp) DNA sequence variation of the trnL-F region, and rooted with Stemona spp., the two Asian species are identified as sister that likely diverged in the Mid-to-Late Pleistocene (0.84–0.13 mya), whereas the divergence of C. pauciflora dates back to the Late Plio-/Pleistocene (<2.6 mya). Phylogeographical analysis of the two East Asian species detected seven cpDNA (trnL-F) haplotypes across 16 populations surveyed, and all of those were fixed for a particular cpDNA haplotype (H_E = 0.0, G_ST = 1). A survey of inter-simple sequence repeats (ISSRs) markers also detected remarkably low levels of within-population diversity (C. japonica: H_E = 0.085; C. heterosepala: H_E = 0.125), and high levels of inter-population differentiation (C. japonica: Φ_ST = 0.736; C. heterosepala: Φ_ST = 0.550), at least partly due to pronounced regional genetic substructure within both species. Non-overlapping distributions of cpDNA haplotypes and strong genetic (cpDNA/ISSR) differentiation among populations and/or regions accord with findings of a nested clade analysis, which inferred allopatric fragmentation as the major process influencing the spatial haplotype distribution of the two species. Based on mismatch distribution analysis and neutrality tests, we do not find evidence of population expansion in both species. Overall, we conclude that components of temperate-deciduous forest types in South Japan and East China are particularly sensitive to range fragmentation, isolation, and enhanced (incipient) species formation through climate-induced expansions of other forest types over glacial and interglacial periods of the (Late) Quaternary.     

Population genetics and conservation of the threatened southeastern beach mouse (Peromyscus polionotus niveiventris): subspecies and evolutionary units

We investigated genetic diversity within the southeastern beach mouse (SEBM-Peromyscus polionotus niveiventris) and also tested the hypothesis that the subspecies recognition of P.p. niveiventris, based on size and color differences, is congruent with this taxon representing a discrete evolutionary lineage. We used ten polymorphic microsatellite loci and mitochondrial cytochrome-b gene DNA sequences to investigate genetic diversity and population structure within the SEBM, and to determine the level of divergence between the SEBM and the nearest known inland subspecies of the oldfield mouse (Peromyscus polionotus rhoadsi). Moderate genetic distances were observed between the SEBM and the inland oldfield mouse based on microsatellite data, with F _ST values ranging from 0.11 to 0.22 between these taxa. Additionally, mitochondrial DNA haplotypes of the SEBM formed a distinct monophyletic group relative to haplotypes sampled from P. p. rhoadsi. Based on previous estimates of rates of mitochondrial DNA evolution in rodents, we inferred that Pleistocene sea-level fluctuations are likely responsible for the historical isolation of the SEBM lineage from mainland P. polionotus. Our data demonstrate the genetic distinctiveness of the SEBM, justifying the current subspecies designation for the SEBM and its continued protection under the United States Endangered Species Act. We classify the Cape Canaveral and Smyrna Dunes Park populations of SEBM as a single evolutionary significant unit. The two known extant allopatric populations of the SEBM showed some differentiation in microsatellite frequencies and were moderately reciprocally distinguishable based on assignment to distinct genetic clusters by a Bayesian admixture procedure. These results justify the classification of these two extant SEBM populations as distinct management units that should be independent targets of management and conservation attention.     

Asynchronous spawning in sympatric populations of a hard coral reveals cryptic species and ancient genetic lineages

Genetic subdivision within a species is a vital component of the evolution of biodiversity. In some species of Acropora corals in Western Australia, conspecific individuals spawn in two seasons 6 months apart, which has the potential to impede gene flow and result in genetic divergence. Genetic comparison of sympatric spring and autumn spawners of Acropora samoensis was conducted to assess the level of reproductive isolation and genetic divergence between the spawning groups based on multiple loci (13 microsatellite loci, the mitochondrial control region and two nuclear introns). Bayesian clustering and principal coordinate analysis of the microsatellite loci showed a high level of genetic differentiation between the spawning groups (F’_ST = 0.30; P < 0.001), as did the sequence data from PaxC and Calmodulin (Φ_ST = 0.97 and 0.31, respectively). At the PaxC locus, the autumn and spring spawners were associated with two divergent lineages that were separated by an evolutionary distance of 1.7% and statistical tests indicate divergent selection in PaxC, suggesting this gene may play a role in coral spawning. This study indicates that the autumn and spring spawners represent two cryptic species, and highlights the importance of asynchronous spawning as a mechanism influencing speciation in corals.     

Microsatellite variation, population structure, and bottlenecks in the threatened copperbelly water snake

Habitat loss and isolation is pervasive in the Midwest U.S. Wetlands are experiencing particularly dramatic declines, yet there is a paucity of information on the genetic impacts of these losses to obligate wetland vertebrates. We quantified the genetic variation of extant populations of a shallow wetland specialist and evaluated potential reductions in population size (i.e. bottlenecks) using seven polymorphic microsatellite DNA markers. We analyzed 228 copperbelly water snakes (Nerodia erythrogaster neglecta), representing populations from three states. Moderate genetic differentiation exists among all three regions (F _ST = 0.12, P < 0.001), with evidence for low levels of differentiation within the federally protected Ohio region (F _ST = 0.025, P = 0.007), and moderate to strong differentiation within the Indiana region (F _ST = 0.23, P < 0.001). Furthermore, Bayesian clustering (i.e. STRUCTURE) supports the separation of the Indiana sites, both from each other and from all other sampling sites. However, it does not support the separation of the Ohio sites from the Kentucky sites. Differentiation among sampling sites did not appear to be related to geographic distance, but rather depended on the quality of terrestrial corridors used for dispersal. Mode shifts in allele frequencies and excess heterozygosity tests were negative, while M-ratio tests were nearly all positive, indicating the likelihood of historical rather than contemporary population bottlenecks. However, potential subspecific intergradation in the Kentucky region may have artificially lowered the M-ratio, and we suggest caution when using the M-ratio approach if intergradation is suspected. Our results have conservation implications for wetland management and management of the copperbelly populations, and emphasizes the importance of protecting wetland complexes.     

Isolation and characterization of 24 microsatellite loci in Paphiopedilum rothschildianum, an endangered slipper orchid

This paper describes the characterization of 24 polymorphic microsatellite loci in Paphiopedilum rothschildianum, an endangered slipper orchid endemic to Sabah, Malaysia. A total of 36 clones containing microsatellite repeat motifs yielded 41 dinucleotide, 20 tetranucleotide and 30 cryptic simple repeat sequences. Thirty microsatellite loci were randomly selected to characterize the population, of which 24 were found to be polymorphic. The expected heterozygosity ranged from 0.0813 to 0.7279 (mean 0.4533) and the observed heterozygosity ranged from 0 to 1 (mean 0.3800). The results indicated that there was limited migration between the three subpopulations and a significant degree of genetic differentiation (F_ST = 0.5098) between populations implying that gene flow has been disrupted as a result of habitat fragmentation. The molecular markers have a potential application in population management, CITES enforcement and the forensic detection of specimens in international trade.     

An exploratory analysis of geographic genetic variation in southern African nyala (Tragelaphus angasii)

We report patterns of genetic variation based on microsatellite, allozyme and mitochondrial control region markers in nyala from geographic locations sampled in South Africa, Mozambique, Malawi and Zimbabwe. Highly significant differences were observed among allele frequencies at three microsatellite loci between populations from KwaZulu-Natal, Limpopo and Malawi, with the Malawi and KwaZulu-Natal groupings showing the highest differentiation (R_ST=0.377). Allozyme frequencies showed minor, non-statistically significant regional differences among the South African populations, with maximum F_ST values of 0.048–0.067. Mitochondrial DNA analyses indicated a unique haplotype in each location sampled. Since none of these indices of population differentiation showed significant correlation to absolute geographic distance, we conclude that geographic variation in this species is probably a function of a distribution pattern stemming from habitat specificity. It is suggested that translocations among geographically distant regional populations be discouraged at present, pending a more elaborate investigation. Transfer of native individuals among local populations may, however, be required for minimizing the likelihood of inbreeding depression developing in small captive populations.     

Microsatellite Variation in Two Populations of Free-Ranging Yellow Baboons (Papio hamadryas cynocephalus)

We investigated genetic variation at six microsatellite (simple sequence repeat) loci in yellow baboons (Papio hamadryas cynocephalus) at two localities: the Tana River Primate Reserve in eastern Kenya and Mikumi National Park, central Tanzania. The six loci (D1S158, D2S144, D4S243, D5S1466, D16S508, and D17S804) were all originally cloned from and characterized in the human genome. These microsatellites are polymorphic in both baboon populations, with the average heterozygosity across loci equal to 0.731 in the Tana River sample and 0.787 in the Mikumi sample. The genetic differentiation between the two populations is substantial. Kolmogornov–Smirnov tests indicate that five of the six loci are significantly different in allele frequencies in the two populations. The mean F _ST across loci is 0.069, and Shriver's measure of genetic distance, which was developed for microsatellite loci (Shriver et al., 1995), is 0.255. This genetic distance is larger than corresponding distances among human populations residing in different continents. We conclude that (a) the arrays of alleles present at these six microsatellite loci in two geographically separated populations of yellow baboons are quite similar, but (b) the two populations exhibit significant differences in allele frequencies. This study illustrates the potential value of human microsatellite loci for analyses of population genetic structure in baboons and suggests that this approach will be useful in studies of other Old World monkeys.     

Morphometric differentiation across House Sparrow Passer domesticus populations in Finland in comparison with the neutral expectation for divergence

To understand the biology of organisms it is important to take into account the evolutionary forces that have acted on their constituent populations. Neutral genetic variation is often assumed to reflect variation in quantitative traits under selection, though with even low neutral divergence there can be substantial differentiation in quantitative genetic variation associated with locally adapted phenotypes. To study the relative roles of natural selection and genetic drift in shaping phenotypic variation, the levels of quantitative divergence based on phenotypes (P_ST) and neutral genetic divergence (F_ST) can be compared. Such a comparison was made between 10 populations of Finnish House Sparrows (n = 238 individuals) collected in 2009 across the whole country. Phenotypic variation in tarsus‐length, wing‐length, bill‐depth, bill‐length and body mass were considered and 13 polymorphic microsatellite loci were analysed to quantify neutral genetic variation. Calculations of P_ST were based on Markov‐Chain Monte Carlo Bayesian estimates of phenotypic variances across and within populations. The robustness of the conclusions of the P_ST–F_ST comparison was evaluated by varying the proportion of variation due to additive genetic effects within and across populations. Our results suggest that body mass is under directional selection, whereas the divergence in other traits does not differ from neutral expectations. These findings suggest candidate traits for considering gene‐based studies of local adaptation. The recognition of locally adapted populations may be of value in the conservation of this declining species.     

Microsatellite DNA Polymorphism in Intensely Enhanced Populations of Sea Trout (Salmo trutta) in the Southern Baltic

Sea trout (Salmo trutta) is an anadromous form of brown trout, a commercially important salmonid species in Europe. Stocking has been used to compensate for the decrease of natural populations and maintenance of fishery and angling catches. Over 1.5 million smolts and 4.5 million alevins are released to Polish coastal rivers each year. Variation at 7 microsatellite loci (Ssa197, Ssa171, Ssa85, Str15, Str73, Str591, and Str543) was used to study genetic polymorphism of spawners returning to 6 rivers. Application of distance method for comparison of pairs of populations based on number of different alleles (F _ST) revealed significant differences between Vistula and Wieprza, and Parseta as well as between Drweca and Wieprza, and Slupia. The level of heterozygosity was similar between most of the studied sea trout populations; considerable differences were found only for Str591. Differences in frequencies of a few alleles between populations were observed. An exact test of sample differentiation based on allele frequencies confirmed lack of significance of differentiation between the 6 pairs of populations (F _ST and R _ST). No admixture was observed in the studied populations. Possible effects of stocking on the genetic polymorphism of the sea trout populations in wild with implications for biotechnology are discussed.     

Population genetic structure of striped mullet, Mugil cephalus, along the coast of China, inferred by AFLP fingerprinting

The striped mullet, Mugil cephalus, is an economically important species for both aquaculture and commercial fisheries in China. In this study, the amplified fragment length polymorphism (AFLP) technique was employed to analyze population genetic diversity and genetic distance between different populations with the aim of elucidating the population structure and gene flow of M. cephalus along the coast of China. A total of 230 fragments with 150–500 bp were identified from 118 individuals by five AFLP primer combinations. The polymorphic loci within populations varied from 46.52% to 64.78%, with an average of 53.91%, and the average heterozygosity from 0.1829 to 0.2282, with an average of 0.2074. The UPGMA phenograms of 118 individuals were constructed based on Dice similarity coefficients and four clusters were recognized. AMOVA analysis revealed that 60.7% of genetic variations were within populations and 39.3% between populations. The estimated genetic distance (φ_ST) value over all polymorphic loci across the six populations was 0.393 (p < 0.0010), indicating a strong population structuring. The pairwise φ_ST value ranged from 0.1112 to 0.5358, with an average of 0.3693. The population pairwise gene flows (average N_m = 0.73) are low. In addition, the result of the Mantel test showed that there was a significant correlation between geographic and genetic distances (r = 0.5434, p = 0.0050). It was speculated that there exist at least four distinct geographic populational subdivisions of M. cephalus along the Chinese coast. This research has provided new molecular data which will aid our understanding of the genetic structure of this species.     

High degree of genetic differentiation in marine three‐spined sticklebacks (Gasterosteus aculeatus)

Populations of widespread marine organisms are typically characterized by a low degree of genetic differentiation in neutral genetic markers, but much less is known about differentiation in genes whose functional roles are associated with specific selection regimes. To uncover possible adaptive population divergence and heterogeneous genomic differentiation in marine three‐spined sticklebacks (Gasterosteus aculeatus), we used a candidate gene‐based genome‐scan approach to analyse variability in 138 microsatellite loci located within/close to (<6 kb) functionally important genes in samples collected from ten geographic locations. The degree of genetic differentiation in markers classified as neutral or under balancing selection—as determined with several outlier detection methods—was low (F_ST = 0.033 or 0.011, respectively), whereas average F_ST for directionally selected markers was significantly higher (F_ST = 0.097). Clustering analyses provided support for genomic and geographic heterogeneity in selection: six genetic clusters were identified based on allele frequency differences in the directionally selected loci, whereas four were identified with the neutral loci. Allelic variation in several loci exhibited significant associations with environmental variables, supporting the conjecture that temperature and salinity, but not optic conditions, are important drivers of adaptive divergence among populations. In general, these results suggest that in spite of the high degree of physical connectivity and gene flow as inferred from neutral marker genes, marine stickleback populations are strongly genetically structured in loci associated with functionally relevant genes.