Are low but statistically significant levels of genetic differentiation in marine fishes 'biologically meaningful'? A case study of coastal Atlantic cod.

A key question in many genetic studies on marine organisms is how to interpret a low but statistically significant level of genetic differentiation. Do such observations reflect a real phenomenon, or are they caused by confounding factors such as unrepresentative sampling or selective forces acting on the marker loci? Further, are low levels of differentiation biologically trivial, or can they represent a meaningful and perhaps important finding? We explored these issues in an empirical study on coastal Atlantic cod, combining temporally replicated genetic samples over a 10-year period with an extensive capture-mark-recapture study of individual mobility and population size. The genetic analyses revealed a pattern of differentiation between the inner part of the fjord and the open skerries area at the fjord entrance. Overall, genetic differentiation was weak (average F(ST) = 0.0037), but nevertheless highly statistical significant and did not depend on particular loci that could be subject to selection. This spatial component dominated over temporal change, and temporal replicates clustered together throughout the 10-year period. Consistent with genetic results, the majority of the recaptured fish were found close to the point of release, with <1% of recaptured individuals dispersing between the inner fjord and outer skerries. We conclude that low levels of genetic differentiation in this marine fish can indeed be biologically meaningful, corresponding to separate, temporally persistent, local populations. We estimated the genetically effective sizes (N(e) ) of the two coastal cod populations to 198 and 542 and found a N(e) /N (spawner) ratio of 0.14.

Detecting genetic structure in migrating bowhead whales off the coast of Barrow, Alaska.

We develop a general framework for analysing and testing genetic structure within a migratory assemblage that is based on measures of genetic differences between individuals. We demonstrate this method using microsatellite DNA data from the Bering-Chukchi-Beaufort stock of bowhead whales (Balaena mysticetus), sampled via Inuit hunting during the spring and autumn migration off Barrow, Alaska. This study includes a number of covariates such as whale ages and the time separation between captures. Applying the method to a sample of 117 bowhead whales, we use permutation methods to test for temporal trends in genetic differences that can be ascribed to age-related effects or to timing of catches during the seasons. The results reveal a pattern with elevated genetic differences among whales caught about a week apart, and are statistically significant for the autumn migration. In contrast, we find no effects of time of birth or age-difference on genetic differences. We discuss possible explanations for the results, including population substructuring, demographic consequences of historical overexploitation, and social structuring during migration.

Cryptic population structuring in Scandinavian lynx: reply to Pamilo.

In a recent Commentary in this journal, Pamilo (2004) criticized our analysis of the spatial genetic structure of the Eurasian lynx in Scandinavia (Rueness et al. 2003). The analyses uncovered a marked geographical differentiation along the Scandinavian peninsula with an apparent linear gradient in the north-south direction. We used computer simulations to check on the proposition that the observed geographical structure could have arisen by genetic drift and isolation by distance in the approximate 25 generations that have passed since the last bottleneck. Pamilo disapproved of our choice of population model and also how we compared the outcome of the simulations with data. As these issues should be of interest to a wider audience we discuss them in some detail.

Cryptic population structure in a large, mobile mammalian predator: the Scandinavian lynx.

The Eurasian lynx (Lynx lynx) is an example of a species that has gone through a severe bottleneck, leading to near extinction in Scandinavia around 1930-- a pattern shared with several other large carnivorous mammals. Here we extend previous genetic analyses of northern European lynx, confirming that lynx from the Scandinavian Peninsula represent a distinct clade differing clearly from European conspecifics. Furthermore, and despite a recent bottleneck and subsequent range expansion, we detect marked genetic differentiation within Scandinavia. This differentiation is largely manifested as a north-south gradient, with a linear increase in the quantity FST/(1 - FST). Aided by computer simulations we find that this pattern is unlikely to have arisen by random genetic drift in the short time since lynx started to expand in the 1950s, suggesting that the spatial structure may predate the bottleneck. Individual-based analyses indicate that, instead of a continuous gradient, Scandinavian lynx may be structured into three more or less distinct groups, possibly corresponding to northern, central and southern subpopulations. The presence of such structuring was unknown previously and was unexpected from general considerations on the mobility of the species, historical data and the absence of geographical barriers. Our study demonstrates how molecular markers may be used to detect cryptic population structure, invisible using traditional methods.

Fine-scaled geographical population structuring in a highly mobile marine species: the Atlantic cod.

Compared with many terrestrial and freshwater environments, dispersal and interbreeding is generally much less restricted in the marine environment. We studied the tendency for a marine species, the Atlantic cod, to be sub-structured into genetically differentiated populations on a fine geographical scale. We selected a coastal area free of any obvious physical barriers and restricted sampling to a 300-km region, well within the dispersal ability of this species. Screening 10 polymorphic microsatellite loci in 6 samples we detected a weak, but consistent, differentiation at all 10 loci. The average FST over loci was small (0.0023) but highly significant statistically, demonstrating that genetically differentiated populations can arise and persist in the absence of physical barriers or great distance. We found no geographical pattern in the genetic differentiation and there was no apparent trend of isolation by distance along the coastline. These findings lend support to the notion that low levels of differentiation are due to passive transport of eggs or larvae by the ocean currents rather than to adult dispersal, the latter being strongly dependent on distance.

Genetic evidence for mixed origin of recolonized sea trout populations.

Anadromous brown trout along the Norwegian Skagerrak coast are genetically differentiated among streams, and there are indications of further substructuring within some streams. Among presumably long-standing populations there is a pattern of increased genetic differentiation with distance, indicating an isolation-by-distance effect. For trout that inhabit streams that have recently been recolonized after the extinction of trout because of acidification, we find evidence for a mixed origin of the recolonizing trout. Both the high levels of gametic phase disequilibrium and the clear deviation from the general pattern of increased genetic differentiation with distance that are seen in recolonized streams, are consistent with recent population admixture, and confirm the loss of the original populations of these acid streams.

Statistical power when testing for genetic differentiation.

A variety of statistical procedures are commonly employed when testing for genetic differentiation. In a typical situation two or more samples of individuals have been genotyped at several gene loci by molecular or biochemical means, and in a first step a statistical test for allele frequency homogeneity is performed at each locus separately, using, e.g. the contingency chi-square test, Fisher's exact test, or some modification thereof. In a second step the results from the separate tests are combined for evaluation of the joint null hypothesis that there is no allele frequency difference at any locus, corresponding to the important case where the samples would be regarded as drawn from the same statistical and, hence, biological population. Presently, there are two conceptually different strategies in use for testing the joint null hypothesis of no difference at any locus. One approach is based on the summation of chi-square statistics over loci. Another method is employed by investigators applying the Bonferroni technique (adjusting the P-value required for rejection to account for the elevated alpha errors when performing multiple tests simultaneously) to test if the heterogeneity observed at any particular locus can be regarded significant when considered separately. Under this approach the joint null hypothesis is rejected if one or more of the component single locus tests is considered significant under the Bonferroni criterion. We used computer simulations to evaluate the statistical power and realized alpha errors of these strategies when evaluating the joint hypothesis after scoring multiple loci. We find that the 'extended' Bonferroni approach generally is associated with low statistical power and should not be applied in the current setting. Further, and contrary to what might be expected, we find that 'exact' tests typically behave poorly when combined in existing procedures for joint hypothesis testing. Thus, while exact tests are generally to be preferred over approximate ones when testing each particular locus, approximate tests such as the traditional chi-square seem preferable when addressing the joint hypothesis.

Spatial structure of lemming populations (Dicrostonyx groenlandicus) fluctuating in density.

The pattern and scale of the genetic structure of populations provides valuable information for the understanding of the spatial ecology of populations, including the spatial aspects of density fluctuations. In the present paper, the genetic structure of periodically fluctuating lemmings (Dicrostonyx groenlandicus) in the Canadian Arctic was analysed using mitochondrial DNA (mtDNA) control region sequences and four nuclear microsatellite loci. Low genetic variability was found in mtDNA, while microsatellite loci were highly variable in all localities, including localities on isolated small islands. For both genetic markers the genetic differentiation was clear among geographical regions but weaker among localities within regions. Such a pattern implies gene flow within regions. Based on theoretical calculations and population census data from a snap-trapping survey, we argue that the observed genetic variability on small islands and the low level of differentiation among these islands cannot be explained without invoking long distance dispersal of lemmings over the sea ice. Such dispersal is unlikely to occur only during population density peaks.

Genetic differentiation among populations of the beetle Bolitophagus reticulatus (Coleoptera: tenebrionidae) in a fragmented and a continuous landscape.

The effect of habitat fragmentation on genetic differentiation among local populations of the fungivorous beetle Bolitophagus reticulatus (Coleoptera: Tenebrionidae) was studied in two contrasting landscapes: one heavily fragmented with forest fragments of variable size surrounded by inhabitable agricultural fields, the other an old forest providing a continuous habitat. The genetic structure of the beetle within each of the two contrasting areas was investigated by means of protein electrophoresis, screening four polymorphic loci in 20 populations from each area. In both areas there were significant genetic differences among local populations, but on average differentiation in the fragmented area was three times greater than in the continuous one, strongly indicating a genetic isolation effect of habitat fragmentation. These genetic results are in accordance with previous studies on dispersal in this species.

Lack of concordance between mtDNA gene flow and population density fluctuations in the bank vole.

The genetic structure of bank voles Clethrionomys glareolus was determined from analyses of mitochondrial DNA (mtDNA) sequences, and compared with previous data on geographical synchrony in population density fluctuations. From 31 sample sites evenly spaced out along a 256-km transect in SE Norway a total of 39 distinct mtDNA haplotypes were found. The geographical distribution of the haplotypes was significantly non-random, and a cladistic analysis of the evolutionary relationship among haplotypes shows that descendant types were typically limited to a single site, whereas the ancestral types were more widely distributed geographically. This geographical distribution pattern of mtDNA haplotypes strongly indicates that the range and amount of female dispersal is severely restricted and insufficient to account for the previously observed synchrony in population density fluctuations. We conclude that geographical synchrony in this species must be caused by factors that are external to the local population, such as e.g. mobile predators.

Demographic genetics of brown trout (Salmo trutta) and estimation of effective population size from temporal change of allele frequencies.

We studied temporal allele frequency shifts over 15 years and estimated the genetically effective size of four natural populations of brown trout (Salmo trutta L.) on the basis of the variation at 14 polymorphic allozyme loci. The allele frequency differences between consecutive cohorts were significant in all four populations. There were no indications of natural selection, and we conclude that random genetic drift is the most likely cause of temporal allele frequency shifts at the loci examined. Effective population sizes were estimated from observed allele frequency shifts among cohorts, taking into consideration the demographic characteristics of each population. The estimated effective sizes of the four populations range from 52 to 480 individuals, and we conclude that the effective size of natural brown trout populations may differ considerably among lakes that are similar in size and other apparent characteristics. In spite of their different effective sizes all four populations have similar levels of genetic variation (average heterozygosity) indicating that excessive loss of genetic variability has been retarded, most likely because of gene flow among neighboring populations.

Temporal allele frequency change and estimation of effective size in populations with overlapping generations.

In this paper we study the process of allele frequency change in finite populations with overlapping generations with the purpose of evaluating the possibility of estimating the effective size from observations of temporal frequency shifts of selectively neutral alleles. Focusing on allele frequency changes between successive cohorts (individuals born in particular years), we show that such changes are not determined by the effective population size alone, as they are when generations are discrete. Rather, in populations with overlapping generations, the amount of temporal allele frequency change is dependent on the age-specific survival and birth rates. Taking this phenomenon into account, we present an estimator for effective size that can be applied to populations with overlapping generations.

Allele frequency estimation at loci with incomplete co-dominant expression.

Although allelic variants at a locus usually are expressed either dominantly or co-dominantly, there are many cases when a gene is dominantly expressed in some individuals and co-dominantly in others. We present a maximum-likelihood procedure for allele frequency estimation in such situations of "incomplete" co-dominant gene expression at an autosomal locus that segregates for two alleles. Our proposed estimator generally is less biased and has a smaller sampling variance than those previously described.