Spatial autocorrelation in biology 2. Some biological implications and four applications of evolutionary and ecological interest

Spatial autocorrelation analysis tests whether the observed value of a variable at one locality is significantly dependent on values of the variable at neighbouring localities. The method was extended by us in an earlier paper to include the computation of correlograms for spatial autocorrelation. These show the autocorrelation coefficient as a function of distance between pairs of localities, and summarize the patterns of geographic variation exhibited by the response surface of any given variable. Identical variation patterns lead to identical correlograms, but different patterns may or may not yield different correlograms. Similarity in the correlograms of different variation patterns suggests similarity in the generating mechanism of the pattern.
The inferences that can be drawn from correlograms are discussed and illustrated. Examination and analysis of variation patterns of several characters or gene frequencies for one population, or of several populations in different places or at different times, permit some conclusions about the nature of the populational processes generating the observed patterns.
Autocorrelation analysis is applied to four biological situations differing in the nature of the data (interval or nominal), in the type of grid connecting the localities (regular or irregular), and the field of application (evolution or ecology). The examples comprise genotypes of individual mice, blood group frequencies in humans, gene frequency variation in a perennial herb, and the distribution of species of trees. The implications of our findings are discussed.     

A simulation study of microevolutionary inferences by spatial autocorrelation analysis

To explore the extent to which microevolutionary inference can be made using spatial autocorrelation analysis of gene frequency surfaces, we simulated sets of surfaces for nine evolutionary scenarios, and subjected spatially-based summary statistics of these to linear discriminant analysis. Scenarios varied the amounts of dispersion, selection, migration, and deme sizes, and included: panmixia, drift, intrusion, and stepping-stone models with 0–2 migrations, 0–2 selection gradients, and migration plus selection. To discover how weak evolutionary forces could be and still allow discrimination, each scenario had both a strong and a weak configuration. Discriminant rules were calculated using one collection of data (the training set) consisting of 250 sets of 15 surfaces for each of the nine scenarios. Misclassification rates were verified against a second, entirely new set of data (the test set) equal in size. Test set misclassification rates for the 20 best discriminating variables ranged from 39.3% (weak) to 3.6% (strong), far lower than the expected rate of 88.9% absent any discriminating ability. Misclassification was highest when discriminating the number of migrational events or the presence or number of selection events. Discrimination of drift and panmixia from the other scenarios was perfect. A subsequent subjective analysis of a subset of the data by one of us yielded comparable, although somewhat higher, misclassification rates. Judging by these results, spatial autocorrelation variables describing sets of gene frequency surfaces permit some microevolutionary inferences.     

Population genetics of the mangrove salt marsh snake, <Emphasis Type="Italic">Nerodia clarkii compressicauda</Emphasis>, in a linear,fragmented habitat

The mangrove salt marsh snake (Nerodia clarkii compressicauda) occupies a unique and disappearing habitat in much of coastal southern Florida. Given extensive habitat fragmentation and high predation pressure in open spaces, it seems likely that populations of N. c. compressicauda consist of isolated groups of related individuals. To assess the degree of population subdivision in this species we genotyped a total of 125 individuals from seven locations along the Florida coast at four microsatellite loci. Overall heterozygosity was moderate (57.7%) and somewhat lower than that seen in other snake species. Population subdivision was particularly pronounced with 19 of 21 sample pair-wise Φ_ST values significantly different from zero and ranging from 0.064 to 0.343 (P ≤ 0.05). About 11 of 39 alleles were private alleles that also tended to be in high frequency in the populations where they occurred (average frequency ~27%). The correlation of genetic and geographic distances was highly significant and positive (r ² = 0.8733 and P < 0.001) with Φ_ST increasing by ~0.01 for every 10 km of separation. Overall, salt marsh snake populations appear to be fractured into isolated neighborhoods on the order of 50–80 km. In spite of its apparent local abundance, we believe that N. c. compressicauda is in need of conservation protection. The combination of extremely low dispersal, narrow habitat requirements, and most importantly, extensive habitat alteration resulting from coastal real estate development may mean that N. c. compressicauda is highly susceptible to population extirpation and potentially extinction.     

Genetic diversity and genotypic differentiation between the sexes in swarm aggregations decrease inbreeding in the Formosan subterranean termite

The life of a colony of subterranean termites, such as Coptotermes formosanus Shiraki (Isoptera: Rhinotermitidae), has natural inbreeding and outbreeding cycles. Reproductives of mature colonies can be replaced by their offspring, which increases the degree of inbreeding in each generation. High degrees of inbreeding may lead to inbreeding depression. In this study we focused on mechanisms for inbreeding avoidance during swarming that do not require kin recognition. We investigated genetic differentiation between swarm aggregations (isolation by distance), genetic diversity within swarm aggregations (multiple colony origin) and genetic differentiation between sexes. Alates were collected from five swarm aggregations in New Orleans, La. The genetic make-up of each swarm aggregation was then described by microsatellite genotyping. Alates from the different swarm aggregations were genetically differentiated; however, no isolation by distance up to at least 1000 m was detected. The dispersal distance of alates was sufficient to guarantee mixing of an average of 13 colonies within swarm aggregations. On average, eleven percent of all possible pairs of alates in each swarm aggregation were putative full siblings. Genotypic frequencies differed significantly between males and females. This could not be explained by sex-biased dispersal. We hypothesize sex-biased investment at the colony level to account for this difference. Genetic differentiation between the sexes and dispersal distances sufficient to promote high genetic diversity within swarm aggregations each facilitate inbreeding avoidance. These observations are consistent with the results of previous studies demonstrating that the majority of simple family colonies in Louisiana populations are headed by unrelated and outbred pairs of reproductives. Received 11 August 2005; revised 5 December 2005; accepted 20 December 2005.     

Isolation by distance in the crustose lichens Candelariella vitellina and Placynthium nigrum colonising gravestones in northeast Scotland

The majority of estimates of isolation by distance are produced byindirect molecular methods rather than direct observation of dispersal events orcolonisation rates. This is in part due to technical difficulties in recordingrare dispersal events in the field, and may be limiting the testing ofmetapopulation models. The fact that tombstones are dated on erection was usedto perform direct measurements of rates of colonisation for two contrasting speciesof crustose lichen, Candelariella vitellina andPlacynthium nigrum, which occur commonly on sandstone andgranite headstones, respectively. The effects of isolation by distancewere observed in the colonisation of sandstone by C. vitellina within the first 50 years of the erection of headstones in thesmallest of three burial grounds investigated in northeast Scotland. When therate of colonisation was estimated over different periods of time the effect ofisolation by distance was seen to decline as the stones aged. There was noapparent effect of isolation by distance on the rate of colonisation of graniteheadstones by P. nigrum at any of the three study sites atany period of colonisation. This study suggests that effects of isolation by distance are likely to behighly case specific, varying between species, between sites and over time. Thisimplies that great care should be taken when producing parameter estimates formetapopulation models.     

Population genetic structure of Sorex unguiculatus and Sorex caecutiens (Soricidae, Mammalia) in Hokkaido, based on microsatellite DNA polymorphism

We investigated the genetic structure of Sorex unguiculatus and Sorex caecutiens populations in Hokkaido, Japan, using hypervariable microsatellite DNA markers. We used five microsatellite loci to type 475 S. unguiculatus individuals from 20 localities on the Hokkaido mainland and four localities from each of four offshore islands (and 11 shrews from one locality in southern Sakhalin for a particular analysis). We used six microsatellite loci to type 240 S. caecutiens individuals from 13 localities on the Hokkaido mainland. Genetic variation was high in mainland populations of both species and low in the island populations of S. unguiculatus. Allelic richness and island size were positively correlated for S. unguiculatus, suggesting that genetic drift occurred on those islands due to small population size. In addition, four insular populations of S. unguiculatus were genetically differentiated from the mainland populations, although clear phylogeographic clustering was not confirmed among populations on the Hokkaido mainland for either S. unguiculatus or S. caecutiens. Heterozygosity excess was observed in more than half of the populations including the mainland populations of the two species, suggesting recent bottleneck events in these populations. Population dynamics of the shrews might be explained by a metapopulation scheme. According to autocorrelation analysis, the extent of non-random spatial genetic structure was approximately 100 km. Isolation by distance was observed in S. unguiculatus, but not in S. caecutiens although there is a positive trend. The lack of correlation for S. caecutiens might have been due to small sample size. Thus, no obvious differences in population genetic structure were found between the two species on the Hokkaido mainland in the present study, while previous investigations using mitochondrial DNA sequences inferred that these two species might have rather different biogeographic histories.     

Surnames and dialects in France: population structure and cultural evolution

To study the isonymy structure of France as related to local language variations, the surname distributions of 6.03 million telephone users registered for the year 2002 were analysed in the 21 conterminous regions, their 94 departments and in 809 towns of the Country. For regions and departments the differences among local dialects were quantified according to the dialecto-metrization of the Atlas Linguistique Fran?ais. We found that Lasker's distance between regions was correlated with geographic distance with r=0.692+/-0.040, while Euclidean (r=0.546+/-0.058) and Nei's (r=0.610+/-0.048) distances were less correlated. Slightly lower correlations were observed for departments. Also, dialectometric distance was correlated with geography (r=0.582+/-0.069 for regions and r=0.617+/-0.015 for departments). The correlations between Lasker and dialectometric matrix distances for regions and departments are r=0.625+/-0.046 and 0.544+/-0.014, respectively, indicating that the common cause generating surname and language diversity accounts for about 35% of the differentiation. Both Lasker and dialectometric distances identify very similar boundaries between Poitou, Centre, Bourgogne and Franche Comptée at the North, and Aquitaine, Limousin, Auvergne, Rh?ne-Alpes in the South. Average Fisher's alpha for France was 7877 the highest value observed for the European countries studied to date. The size of alpha in most French towns indicates considerable recent immigration.     

Surnames in Western Europe: a comparison of the subcontinental populations through isonymy

We studied the isonymic structure of Western Europe using the distributions of 26.2 million surnames in 8 countries, 125 regions and 2094 towns of the Subcontinent. We found that, for the whole of Western Europe, Nei's distance was correlated with geographic distance (r=0.610+/-0.009). It was observed that at long geographic distances the isonymyc distance stays below linearity and tends to become asymptotic, and this was attributed to long distance migration. A dendrogram of the125 regions was built and the clusters identified by the dendrogram are almost exactly coincident with the nations of the Subcontinent. Random inbreeding calculated from isonymy, F(ST), was highest in Spanish regions, and lowest in France. The geographical distribution of alpha in 2094 towns, high in the Center and East of the Subcontinent and lower in Spain, is compatible with the settlement of subsequent waves of migrants moving from the West and from the South toward the centre of the Continent. The present surname structure of Western Europe is strictly linked to local languages.     

Ring species – Do they exist in birds?

According to current theory, the splitting of a single species into two is best observed by a ring of intergrading populations which occupy a ring-shaped distribution area and whose terminal populations not only meet but overlap and co-occur without or with only little hybridization. The three most discussed examples in birds are revisited here. The great tit complex (Parus major s. l.) turned out to be an assemblage of four subspecies groups forming a secondary ring of population. The herring gull/lesser black-backed gull complex (Larus argentatus s. l.) forms a circumpolar circle of intergrading populations, but lacks the crucial cornerstone, the geographical overlap. The greenish warbler complex (Phylloscopus trochiloides s. l.) is close to the ideal ring species, but the ring is not complete and the Siberian zone of overlap needs further clarification and characterisation.     

Allozyme data reveal genetic diversity and isolation by distance in sympatric Glyphidrilus Horst, 1889 (Oligochaeta: Almidae) of the Lower Mekong River Basin

The genus Glyphidrilus, comprised of semi-aquatic freshwater earthworms that live in an ecotone between terrestrial and freshwater ecosystems, are widely distributed along riverbanks and rice paddy systems. Two Glyphidrilus species (Glyphidrilus vangviengensis and Glyphidrilus mekongensis) are endemic in the Lower Mekong River Basin and are sympatric from Northern Thailand to Southern Laos. However, species delimitation among the Mekong Glyphidrilus remains unclear because the key morphological traits in semi-aquatic earthworms are highly polymorphic. This study assessed the distinction between G. vangviengensis and G. mekongensis using allozyme electrophoresis. A total of 752 individuals collected from 33 localities were screened for 10 putative loci from seven enzymatic systems, revealing that G. vangviengensis and G. mekongensis are two distinct species, according to their different allelic patterns and high genetic distance. A low genetic differentiation within each species was indicated by the pairwise Nei's D and F_ST analyses, and the absence of population structure was detected by AMOVA and Bayesian structure analyses. However, a significant isolation by distance, but not vicariance, was observed, which is probably due to the river current causing translocation downstream and so gene flow between adjacent localities. The genetic diversity of the Mekong Glyphidrilus was relatively high and comparable to other earthworm taxa, and several localities showed deviation from Hardy–Weinberg equilibrium. An additional cryptic species from Ban Hat Khamphi, Loei, Thailand, was inferred.     

Spatial autocorrelation and linkage of Mendelian RAPD markers in a population of Picea abies Karst

The spatial clustering of single- and di-locus genotypes in a natural, continuous population of Norway spruce was investigated using 69 Mendelian Random Amplified Polymorphic DNA (RAPD) markers that covered about 15 of the species genome, and whose linkage relationships were known. Spatial autocorrelation techniques and randomization tests, applied to both single- and di-locus genotypes, revealed a weak, though significant, spatial structure at the scale 0-200 m (5 of single-locus and 7 of di-locus genotypes). To assess the relative importance of isolation by distance and linkage between markers on their spatial genetic structuring, we grouped joins between sampled trees into equivalence categories expected to show similar, specific patterns of spatial distribution under isolation by distance. Results from both single- and di-locus analyses were consistent with the existence of patches of like homozygotes (about 8 and 11 of loci at the single- and di-locus level, respectively) surrounded by a mix of like heterozygotes. Similar structuring has been predicted by simulation models under isolation by distance and selective neutrality. Overall, linkage between markers accounted for an increase of spatial clumping of di-locus genotypes involving tightly linked loci with recombination fractions up to 0.1, a consequence of limited, stochastic spread of single-locus genotypes in space. Our results support the hypothesis that isolation by distance and linkage have a small, though significant, effect even within continuous forest tree populations. In general, the spatial distribution of multilocus genotypes within populations should be interpreted with caution when linkage relationships among the markers used are unknown.     

Differential population structuring and demographic history of two closely related fish species, Japanese sea bass (Lateolabrax japonicus) and spotted sea bass (Lateolabrax maculatus) in Northwestern Pacific

The Quaternary cold periods in the Northwestern Pacific are thought to have heavily influenced the amount and distribution of intraspecific genetic variation in marine fishes. To estimate the demographic history and genetic structure of Lateolabrax maculatus and L. japonicus in the Northwestern Pacific, 256 individuals were sampled from 19 localities throughout the distribution range of the two species. Mitochondrial DNA variation was analyzed using DNA sequence data from the cytochrome b gene and control region. Nucleotide diversity was much higher in L. japonicus (0.030) than in L. maculatus (0.012). The demographic history of the two species was examined using neutrality tests and mismatch distribution analyses and results indicated Pleistocene population expansion in both species. Estimates of population expansion time suggested earlier population expansion in L. japonicus than in L. maculatus. Molecular variance analyses showed differential genetic structuring for these two closely related species. The results indicated that L. japonicus is panmictic throughout its range. In contrast, populations of L. maculatus showed statistically significant levels of genetic structuring. Pattern of isolation by distance was observed in L. maculatus, suggesting that L. maculatus is in genetic equilibrium. In contrast, L. japonicus did not exhibit isolation by distance.     

Seed banks cause elevated generation times and effective population sizes of Arabidopsis thaliana in northern Europe

It is commonly found that effective population sizes of natural populations are much smaller than census sizes of plants and animals. However, theoretical studies have shown that factors rarely investigated empirically, like seed banks in plants and diapause in animals, may have profound influence on effective sizes. Here we investigate whether the presence of seed banks can explain the relatively high genetic variability observed in northern European Arabidopsis thaliana populations with small census sizes. We have genotyped three above- and below- ground cohorts in 27 Norwegian populations using single nucleotide polymorphism markers. Although the populations varied extensively in levels of variability within and between cohorts, standard genetic population measures were comparable to those obtained in previous studies on above-ground cohorts using microsatellite markers. Estimated effective population sizes are larger for total populations (containing both seed bank and above-ground cohorts for 1 year) compared to each of the cohorts considered separately. Using a conservative approach, we find that the effective sizes are larger than census sizes of local populations, and that the effective generation time is higher than 1 year (3–4 years, on average), making A. thaliana a perennial semelparous plant at many northern European localities.     

Strong gene flow and lack of stable population structure in the face of rapid adaptation to local temperature in a spring-spawning salmonid, the European grayling (Thymallus thymallus)

Gene flow has the potential to both constrain and facilitate adaptation to local environmental conditions. The early stages of population divergence can be unstable because of fluctuating levels of gene flow. Investigating temporal variation in gene flow during the initial stages of population divergence can therefore provide insights to the role of gene flow in adaptive evolution. Since the recent colonization of Lake Lesjaskogsvatnet in Norway by European grayling (Thymallus thymallus), local populations have been established in over 20 tributaries. Multiple founder events appear to have resulted in reduced neutral variation. Nevertheless, there is evidence for local adaptation in early life-history traits to different temperature regimes. In this study, microsatellite data from almost a decade of sampling were assessed to infer population structuring and its temporal stability. Several alternative analyses indicated that spatial variation explained 2-3 times more of the divergence in the system than temporal variation. Over all samples and years, there was a significant correlation between genetic and geographic distance. However, decomposed pairwise regression analysis revealed differing patterns of genetic structure among local populations and indicated that migration outweighs genetic drift in the majority of populations. In addition, isolation by distance was observable in only three of the six years, and signals of population bottlenecks were observed in the majority of samples. Combined, the results suggest that habitat-specific adaptation in this system has preceded the development of consistent population substructuring in the face of high levels of gene flow from divergent environments.     

Genetic variation and population structure of marbled sole Pleuronectes yokohamae and cresthead flounder P. schrenki in Japan inferred from mitochondrial DNA analysis

Nucleotide sequence analysis in the first half of the mitochondrial control region was used to estimate genetic variation and population structure of marbled sole Pleuronectes yokohamae and cresthead flounder Pleuronectes schrenki, obtained from 10 localities in Japan. Overall estimate of pairwise population F_ST and NJ tree suggested genetic distinction between P. yokohamae and P. schrenki. SAMOVA also demonstrated two distinct genetic groups of the samples, MS (marbled sole, Honshu and southernmost Hokkaido) and CF(cresthead flounder, Hokkaido), which was pertinent to the natural range of P. yokohamae and P. schrenki, respectively. However, the haplotypes common to P. yokohamae and P. schrenki were found in all four samples of the CF group and two of six samples of the MS group. Haplotype and nucleotide diversities were high within MS and CF groups, and the mismatch distribution analysis showed a bimodal profile in both groups, suggesting secondary contact between allopatrically evolved MS and CF groups. Thus, the observed haplotype sharing between two species might have been the consequence of interspecific hybridization after their divergence. In addition, F_ST estimates and Mantel test suggested that the population structure of P. yokohamae might have been influenced by low or restricted gene flow with isolation by distance, whereas no such population structuring was implied for P. schrenki. The observed difference in population structuring between two species was surmised by the difference of migration character.     

Genetic divergence and isolation by distance in the West Atlantic sea anemone Actinia bermudensis (McMurrich, 1889)

The sea anemone Actinia bermudensis [Proc. Acad. Nat. Sci. Philadelphia 41 (1889) 102] has a widespread distribution over the Western Atlantic, where it has been reported from Bermuda and Bahamas to southern Brazil. Nonetheless, reproductive features of this species, which broods asexually produced offspring, suggest that it has low dispersal capabilities. This raises a question about the specific status of the Brazilian population of A. bermudensis. The aim of this paper was to evaluate the taxonomic status of the Brazilian population, by genetically comparing it with the type locality (Bermuda) population. For this, we used allozyme electrophoresis to analyze samples from both populations. Thirteen enzyme systems (16 loci) yielded interpretable results for all populations. A high differentiation was encountered for the Western Atlantic A. bermudensis population (F_ST=0.434), partially due to the presence of a diagnostic locus (LAP) between Bermuda and Brazilian populations. This result could be taken as an indication that the Brazilian population should be elevated to a specific status. We do expect some genetic differentiation, however, between natural populations separated over wide geographic distances. Finally, since Mantel and regression analyses did not rule out the isolation by distance model, we have conservatively decided to maintain the specific status of A. bermudensis from Brazil.     

Effects of sample size,number of markers,and allelic richness on the detection of spatial genetic pattern

The influence of study design on the ability to detect the effects of landscape pattern on gene flow is one of the most pressing methodological gaps in landscape genetic research. To investigate the effect of study design on landscape genetics inference, we used a spatially‐explicit, individual‐based program to simulate gene flow in a spatially continuous population inhabiting a landscape with gradual spatial changes in resistance to movement. We simulated a wide range of combinations of number of loci, number of alleles per locus and number of individuals sampled from the population. We assessed how these three aspects of study design influenced the statistical power to successfully identify the generating process among competing hypotheses of isolation‐by‐distance, isolation‐by‐barrier, and isolation‐by‐landscape resistance using a causal modelling approach with partial Mantel tests. We modelled the statistical power to identify the generating process as a response surface for equilibrium and non‐equilibrium conditions after introduction of isolation‐by‐landscape resistance. All three variables (loci, alleles and sampled individuals) affect the power of causal modelling, but to different degrees. Stronger partial Mantel r correlations between landscape distances and genetic distances were found when more loci were used and when loci were more variable, which makes comparisons of effect size between studies difficult. Number of individuals did not affect the accuracy through mean equilibrium partial Mantel r, but larger samples decreased the uncertainty (increasing the precision) of equilibrium partial Mantel r estimates. We conclude that amplifying more (and more variable) loci is likely to increase the power of landscape genetic inferences more than increasing number of individuals.     

Species-Specific Population Structure in Rock-Specialized Sympatric Cichlid Species in Lake Tanganyika,East Africa

Species richness and geographical phenotypic variation in East African lacustrine cichlids are often correlated with ecological specializations and limited dispersal. This study compares mitochondrial and microsatellite genetic diversity and structure among three sympatric rock-dwelling cichlids of Lake Tanganyika, Eretmodus cyanostictus, Tropheus moorii, and Ophthalmotilapia ventralis. The species represent three endemic, phylogenetically distinct tribes (Eretmodini, Tropheini, and Ectodini), and display divergent ecomorphological and behavioral specialization. Sample locations span both continuous, rocky shoreline and a potential dispersal barrier in the form of a muddy bay. High genetic diversity and population differentiation were detected in T. moorii and E. cyanostictus, whereas much lower variation and structure were found in O. ventralis. In particular, while a 7-km-wide muddy bay curtails dispersal in all three species to a similar extent, gene flow along mostly continuous habitat appeared to be controlled by distance in E. cyanostictus, further restricted by site philopatry and/or minor habitat discontinuities in T. moorii, and unrestrained in O. ventralis. In contrast to the general pattern of high gene flow along continuous shorelines in rock-dwelling cichlids of Lake Malawi, our study identifies differences in population structure among stenotopic Lake Tanganyika species. The amount of genetic differentiation among populations was not related to the degree of geographical variation of body color, especially since more phenotypic variation is observed in O. ventralis than in the genetically highly structured E. cyanostictus. [Reviewing Editor: Dr. Rafael Zardoya]     

Population structure and anthropometric variation in rural western Ireland: Isolation by distance and analysis of the residuals

The study of human population structure allows the assessment of cultural and historical influences on mating probabilities, and, hence, genetic variation. A commonly used model is isolation by distance, which predicts a negative exponential relationship between genetic similarity and geographic distance. Anthropometric data collected during the 1930's for 261 adult women in 12 towns of rural western Ireland were used to test the isolation by distance model and to assess the influence of cultural factors upon the fit of the model. The effects of recent migration were tested by using two additional data subsets, one excluding known intercounty migrants and the other consisting of unmarried women, only in an attempt to control partially for local migration upon marriage. Deviations from the expected isolation by distance model were analyzed using rotational fitting and regression analysis. Estimates of the isolation by distance parameters agree closely with independent estimates from isonymy and with estimates obtained in other studies of rural European population structure. Analysis of the residuals indicates three major factors which contribute to deviations from the expected model: recent migration upon marriage, age variation among groups, and variation in population size and/or transportation opportunities. Variation in population size was tested using the gravity model of economic geography by regressing the residuals from the isolation by distance model for each pair of towns on the product of their population sizes. The best fit occurred for the unmarried sample, as expected from ethnographic evidence, since rural–urban migration was most common among unmarried women.