Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Genetic Structure in a Small Pelagic Fish Coincides with a Marine Protected Area: Seascape Genetics in Patagonian Fjords

  • Cristian B. Canales-Aguirre ,

    cristian.canales@ulagos.cl

    Affiliations Laboratorio de Genética y Acuicultura, Departamento de Oceanografía, Facultad de CienciasNaturales y Oceanográficas, Universidad de Concepción, Concepción, Casilla 160-C, Chile, Laboratorio de EcologíaEvolutiva y Filoinformática, Departamento de Zoología, Facultad de CienciasNaturales y Oceanográficas, Universidad de Concepción, Concepción, Casilla 160-C, Chile, Centro i~mar, Universidad de Los Lagos, Camino a Chinquihue 6 km, Puerto Montt, Chile

    ORCID http://orcid.org/0000-0002-8468-6139

  • Sandra Ferrada-Fuentes,

    Affiliations Laboratorio de Genética y Acuicultura, Departamento de Oceanografía, Facultad de CienciasNaturales y Oceanográficas, Universidad de Concepción, Concepción, Casilla 160-C, Chile, Laboratorio de EcologíaEvolutiva y Filoinformática, Departamento de Zoología, Facultad de CienciasNaturales y Oceanográficas, Universidad de Concepción, Concepción, Casilla 160-C, Chile

  • Ricardo Galleguillos,

    Affiliation Laboratorio de Genética y Acuicultura, Departamento de Oceanografía, Facultad de CienciasNaturales y Oceanográficas, Universidad de Concepción, Concepción, Casilla 160-C, Chile

  • Cristián E. Hernández

    Affiliation Centro i~mar, Universidad de Los Lagos, Camino a Chinquihue 6 km, Puerto Montt, Chile

Genetic Structure in a Small Pelagic Fish Coincides with a Marine Protected Area: Seascape Genetics in Patagonian Fjords

  • Cristian B. Canales-Aguirre, 
  • Sandra Ferrada-Fuentes, 
  • Ricardo Galleguillos, 
  • Cristián E. Hernández
PLOS
x

Abstract

Marine environmental variables can play an important role in promoting population genetic differentiation in marine organisms. Although fjord ecosystems have attracted much attention due to the great oscillation of environmental variables that produce heterogeneous habitats, species inhabiting this kind of ecosystem have received less attention. In this study, we used Sprattus fuegensis, a small pelagic species that populates the inner waters of the continental shelf, channels and fjords of Chilean Patagonia and Argentina, as a model species to test whether environmental variables of fjords relate to population genetic structure. A total of 282 individuals were analyzed from Chilean Patagonia with eight microsatellite loci. Bayesian and non-Bayesian analyses were conducted to describe the genetic variability of S. fuegensis and whether it shows spatial genetic structure. Results showed two well-differentiated genetic clusters along the Chilean Patagonia distribution (i.e. inside the embayment area called TicToc, and the rest of the fjords), but no spatial isolation by distance (IBD) pattern was found with a Mantel test analysis. Temperature and nitrate were correlated to the expected heterozygosities and explained the allelic frequency variation of data in the redundancy analyses. These results suggest that the singular genetic differences found in S. fuegensis from inside TicToc Bay (East of the Corcovado Gulf) are the result of larvae retention bya combination of oceanographic mesoscale processes (i.e. the west wind drift current reaches the continental shelf exactly in this zone), and the local geographical configuration (i.e. embayment area, islands, archipelagos). We propose that these features generated an isolated area in the Patagonian fjords that promoted genetic differentiation by drift and a singular biodiversity, adding support to the existence of the largest marine protected area (MPA) of continental Chile, which is the Tic-Toc MPA.

Introduction

Marine environmental landscape parameters play an important role in promoting population genetic differentiation in marine organisms [1]. Consequently, identifying environmental parameters that promote population genetic differentiation is a major focus of study in evolutionary biology [1]. Most research on the effects of the environmental marine landscape on the genetics of population structure has been qualitative (e.g., [2,3]). However, qualitative research may not always be completely successful in identifying the factors that are responsible for the observed genetic structure of natural populations, and most importantly, they do not evaluate those environmental factors explicitly. In fact, few studies evaluate both: genetic and marine environmental data [3]. Manel et al. [4] introduced the concept of landscape genetics, which is able to explain spatial genetic patterns through landscape features (i.e. geographic, physic and chemical variables) and spatial statistics [4,5]. To date, most studies that used this approach have been performed in terrestrial organisms, leaving marine and freshwater organisms mostly unexplored [6]. Recently, concepts such as seascape genetics or marine landscape genetics have started to appear in studies that evaluate how biotic and abiotic factors promote microevolutionary processes in marine species (i.e. fishes, mollusks, crustaceans [1,3,7]). Although different marine habitats (i.e. estuary, open sea, intertidal, pelagic, benthic) could potentially affect the genetic diversity within species, fjord habitats in particular have the potential to greatly affect population genetic diversity due to the complex scenario produced by their heterogeneous geography and environmental characteristics.

Fjords are deep, high-latitude estuaries at have been excavated or modified by glaciers [812]. These estuaries are productive ecosystems that connect the open sea with freshwater from land drainage and melting ice [12,13]. In addition, this ecosystem has been characterized mainly by strong fluctuations in salinity, temperature, pH, oxygen [14] and ocean circulation patterns [15] such as mesoscale eddies and fronts [16]. These environmental characteristics have been indicated as drivers of population differentiation [2,1620]. For example, there is evidence of the effect of environmental oscillations on the marine organisms of fjords at different levels of organization: changes in composition of macrobenthic and zooplankton communities [2123], differences in mortality and growth [24,25], abundance and search efficiency [26]. Environmental factors associated with fjords (i.e. temperature, salinity, oxygen, pH, and nutrients) have been proposed as causes of trophic and reproductive adaptation [2729], and transport and retention of larvae [30,31]. Also, other studies have found population genetics differentiation between inner and outer fjords waters [25,32,33]. In such cases, oceanographic features can be a barrier to dispersal at different ontogenetic stages, by restricting gene flow and increasing intraspecific divergence.

The Chilean Patagonian fjords constitute one of the largest fjord regions in the world, extending from latitude 41.5°S (Reloncaví Fjord) to latitude 55.9°S (Cape Horn) and covering a total of 240,000 km2 [12]. The geographic landscape of this region includes channels, estuaries, archipelagos, fjords, bays, peninsulas and islands [12]. In addition, this ecosystem has been characterized mainly by strong fluctuations in salinity, temperature, pH, oxygen [14] and circulation patterns [15]. The Patagonian sprat Sprattus fuegensis is a small pelagic marine fish of economic importance that inhabits from latitude 41°S, specifically in inner waters and fjords in the south of Chile to latitude 40°S in Argentina, including the Falkland Islands [3438]. It lives a maximum of 6 years [39] and it is a partial spawner [38,4042]. Female sprats mature at an average length of 13.5 cm [38] and produce pelagiceggs and larvae [13,43,44]. Thefirst developmental stages of S. fuegensis are mainly abundant in the inner waters of Chiloé Island, channels and fjords in Chile [13,43,44], and in the Atlantic Ocean they have been reported near Santa Cruz, Argentina and southward to the Falkland Islands [34,45]. We used S. fuegensis as a model to investigate how environment can shape the genetics structure of populations because: (1) it inhabits fjords and channels which have been shown to have high environmental oscillations and in general are habitats with low levels of pollution [12,14,15]; (2) it inhabits mainly the first 50 m of the water column [3438] where environmental variables show high oscillations (see [4648]); (3) there are no studies that evaluate seascape genetics in a fish that lives in fjords and channels in the Southern Hemisphere; (4) its geographic distribution is not only restricted to fjords and channels but extends further north into Argentina [3438], allowing for further comparisons of genetics structure between homogeneous and heterogeneous environment; and (5) the species is economically important in Chile and further understanding of the structure of its populations will be useful in the management and conservation of stocks [49].

Based on the characteristics of the Chilean fjord, we propose that the small pelagic fish S. fuegensis has a large population genetic differentiation promoted by local fjord conditions. Given the geography of the area, we expected to find at least two genetic clusters: one group from the north of Chilean Patagonia (i.e. inner water of Chiloé (~42°-43°S)and fjords close to Aysén (~45°S), and another group in the most distant locality of the Strait of Magellan (~53° S). To test this hypothesis, we genotyped 282 adult S. fuegensis that were collected in 10 locations using eight species-specific microsatellites. We described the genetic diversity and population structure of S. fuegensis along the Chilean fjords and we evaluated the effect of marine environmental variables (i.e. temperature, salinity, oxygen, pH, nutrients, and ocean circulation pattern) that are related to the causal mechanisms (i.e. gene flow, genetic drift) of a population structure. Based on non-Bayesian and Bayesian approaches we found a strong genetic structure in this species, which is correlated with temperature ranges and nitrate concentration, two factors that could be affecting local productivity, growth rates and therefore population dynamics. Finally, we discuss how the oceanographic landscape can promote this divergence and how our results support the existence of the Marine Protected Area (MPA) located in this area (i.e. Tic-Toc MPA).

Materials and Methods

Sample collection

A total of 282 individuals were collected from ten locations in the Chilean Patagonian fjords (Fig 1), including the inner sea of Chiloé and the particular fjords where S. fuegensis has been recorded. Locations were selected based on early studies from scientific cruises [13,50] between latitudes 41° and 46°S, except the most southern location (i.e. 53°S), which was selected based on personal communications with artisanal fishermen from Punta Arenas (i.e. in gathering information about catching sites of S. fuegensis in order to focus the sampling efforts in this area). In addition, samples collected in this southern location (i.e. 53°S) are important to the study questions given that the southern tip of South America is significantly different in temperature, phosphate and nitrate from the habitats inhabited by the northern populations (e.g. 41°S) [4648,5156]. Moreover, different types of vertical structures have been described for the Chilean austral channels and fjords, according to: temperature (i.e. 11°C), salinity (i.e. 7psu), dissolved oxygen-pH (i.e. 5 ml/L-1), phosphate-nitrate (i.e. 7 μM), and silicate (i.e. 9 μM), which support the differences among physical and chemical characteristics of the water column from Chilean Patagonia [57,58]. We sampled adults during the S. fuegensis spawning season (September and December) because this season represents the most robust period for delineating population genetic structure [59]. Sampling within this season avoids including juveniles or earlier life history stages, therefore preventing overestimating genetic differentiation by the presence of close relatives (Allendorf-Phelps effect [59,60]). In this way, we tested the null hypothesis that individuals were randomly assorted in order to spawn in different spawning areas [59,61]. Muscular tissue obtained from fillet from each individual was sampled and stored in 96% ethanol for further analyses.

thumbnail
Fig 1. Map showing sample locations (red dots) of Sprattus fuegensis.

Small map shows places named in the main text. The Tic-Toc MPA is shown in the green area.

https://doi.org/10.1371/journal.pone.0160670.g001

Environmental database

The currently available global marine environmental databases (e.g. BioOracle, AquaMaps, MARSPEC) have large gaps in information related to the inner sea of Chiloé and the fjords. We therefore compiled an environmental marine database based on the published literature and oceanographic research cruises. These records were obtained from the CIMAR-FIORDOS oceanographic research program conducted between 1995–2006 (http://www.shoa.cl/n_cendhoc/productos/reporte_datos.php), whose goal was to compile oceanographic and biological information from the inner sea of Chiloé, channels, estuaries and fjords from the Chilean Patagonia using the same cruise routes in different years but in similar seasons (i.e. spring—summer). The oceanographic environmental measures from the CIMAR-FIORDOS program used in this study were temperature, salinity, pH, oxygen, phosphate, and nitrate. We obtained data from the CIMAR-Fiords program for each variable, from nine different depth of the water column (i.e. 0, 2, 5, 10, 15, 20, 25, 50, and 100) in each of the 10 sampling locations. Then, we estimated the maximum, minimum, average and range in each location for the above-mentioned six marine environmental variables, in order to capture the range of conditions experienced by this species. We show the detailed information on cruises, seasons, year and references in S1 Table [46,47,5156,62].

Genetic database

Total genomic DNA was isolated using NucleoSpin tissue Kit (Machery-Nagel) and carried out according to the manufacturers' recommendations. The quality and quantity of DNA purification were measured in an Eppendorf biophotometer (Eppendorf AG, Hamburg, Germany) and the template DNA was diluted to 20 ng/μL for the polymerase chain reaction (PCR) amplifications. We used eight tetranucleotide microsatellites with loci described for S. fuegensis by Ferrada-Fuentes et al. [63] (i.e. Spfu_6, Spfu_9, Spfu_29, Spfu_30, Spfu_42, Spfu_44, Spfu_45, and Spfu_48). These loci were amplified following the protocol described previously by Ferrada-Fuentes et al. [63] and in the PCR procedures we included both positive and negative controls. The PCR products were run on an ABI-3130xl sequencer and sized with Naurox size standard. This was performed by locus in order to avoid a likely bias that might be generated via analyses per location. Results were analyzed using GeneMapper version 3.7 (Applied Biosystems).

Pre-processing genetics dataset

Because large samples are expected to have more alleles than small samples, and the number of individuals per locality was not homogeneous, we conducted a rarefaction analysis to estimate how many individuals we would need in order to detect all alleles present in a population (i.e. allelic richness, AR) in HP-RARE [64,65]. Outputs of allelic richness obtained from rarefaction analyses indicated that the average expected number of alleles in our within-location standardized sample size (i.e. n = 12) was less than our smallest sample size obtained in the field (i.e. n = 24 to Zone_H and Zone_L), and by clusters (i.e. n = 16) was also smaller than the cluster with the lowest number of individuals sampled (i.e. n = 28 to Cluster in Zone D), therefore our number of individuals was well-suited to further analysis (Table 1).

thumbnail
Table 1. Genetic diversity parameters per sampling location and genetic cluster in microsatellite loci of Sprattus fuegensis.

https://doi.org/10.1371/journal.pone.0160670.t001

To evaluate the quality of the genetic database, we estimated the presence of genotyping errors such as drop-out alleles, stutter bands, and possible presence of null alleles. These analyses were conducted in the MICRO-CHECKER v2.2.3 software [66]. According to MICRO-CHECKER, several loci showed that the general excess of homozygotes is distributed across most allele size classes yielding possible deviations from Hardy-Weinberg equilibrium and presence of null alleles [66]. Taking into account that the presence of null alleles could have an impact on the estimation of population differentiation [67], and in order to avoid a decrease in power in further analyses [68], we employed model-based clustering and Bayesian assignment methods [6971]. These methods take into account null alleles and significantly improve their accuracy in GENELAND software [70]. In addition, simulations including datasets that include the presence of null alleles have demonstrated that genetic clustering outputs do not show more gene pools than there are in reality [72] and that they improve significantly their precision in determining genetic clusters [70]. Therefore, we used the raw microsatellite dataset without any correction for null alleles to infer the number of population clusters. Then, in order to be more conservative and to achieve congruent outcomes, we tested again each locus at a time in order to remove the possible null allele effect.

Finally, in order to avoid inflating patterns of genetic structure due to sibship control (i.e. effect of sampling families [7375]), we ruled out putative full-sibs within samples for each location. To identify full sibs we use the maximum-likelihood method implemented in COLONY v2.0.0.1 [7678]. Full sib analysis was conducted using the ‘long length of run’ and ‘high likelihood precision’ options implemented in COLONY. The outcome from the full sibs identification analysis did not show putative full sibs in the data set, therefore we continued with further analyses without excluding any individuals from each location.

Genetic variability

The total number of alleles (NA), expected (HE) and observed (HO) heterozygosity were estimated to determine the genetic variability of the samples; these parameters were calculated for each locus and locality using GENALEX v6.5 software [79]. To determine whether localities had significant deviations to the Hardy-Weinberg equilibrium and linkage-disequilibrium we conducted analyses in ARLEQUIN v3.1 [80] and GENEPOP 3.1 [81,82], respectively. Pairwise FST and RST comparisons between sampling locations were obtained from ARLEQUIN where the p-value was obtained after 10,100 permutations. In addition, two standardized measures of genetic differentiation were included in order to infer demographic processes such as genetic drift and migration on genetic population structure, as suggested by Meirmans and Hedrick [83]. Sequential Bonferroni correction [84] for multiple comparisons was applied when necessary.

Number of genetic clusters

To infer genetic cluster number (K) in our sample set, we used two Bayesian approaches based on the clustering method which differed in that they: a) incorporated or not a null allele model, and b) useda non-spatial or spatial algorithm. We selected this approach because Bayesian models capture genetic population structure by describing the genetic variation in each population using a separate joint posterior probability distribution over loci, therefore they incorporate uncertainty into the analyses. We used STRUCTURE v.2.3.3 [85,86], which does not incorporate a null allele model, but uses a non-spatial model based on a clustering method and it is able to quantify the individual genome proportion from each inferred population. A previous run had been carried out to define what ancestry models (i.e. no admixture model and admixture model) and allele frequency models (i.e. correlated and uncorrelated allele frequency models) fit our dataset. All these previous runs were conducted with locality information prior to improving the detection of structures when these could be weak [87]. The parameters of previous simulations included five runs with 50,000 iterations following a burn-in period of 5,000 iterations for K = 1–10 as number of tested clusters. Before choosing models with which to run our dataset, we evaluated Evanno’s index ΔK [88] to identify whether different models yielded different K values, implemented in STRUCTURE HARVESTER [89]. Finally, to choose the best model to run our data we comparedthe marginal likelihood of each model which was evaluated using Bayes Factor (BF). The best two models were the no admixture and correlated frequency allele, in which sampling locations were setas an informative prior (S2 Table). These models were used in further analyses. Because the admixture model could have more biological sense, was also tested the analyses with it, but it did not yield significantly different results. The final simulations were run, testing k = 2, with 500,000 iterations of burn-in and a run length of 1,000,000 and all these were replicated ten times independently. Then, we used GENELAND v.0.3 [90], which incorporates geographic information (i.e. coordinates) in a spatial model in order to detect spatial discontinuities among populations with possible uncertainty in spatial coordinates [91] and a null allele model that improves significantly their accuracy to inferences [70]. Similarly to the methods with STRUCTURE as described above, we ran short analyses to determine what model (i.e. correlated or uncorrelated frequency models) would best fit our dataset. All runs were performed using the “null allele model” setting, given that it may have been present in our data. Previous simulations were run for testing K = 1–10, using 1,000,000 Markov chain Monte Carlo (MCMC) iterations, with a thinning interval of 10,000 and all those runs were replicated five times. The selection of the best model was evaluated using BF. The best model used was the correlated frequency model (S2 Table). The final simulations were run testing K = 2, using 10,000,000 MCMC iterations with a thinning interval of 10,000 and all those that were run were replicated ten times each. To identify the number of genetic clusters present in our data we made a graphic with density probability, per each K, per iteration. Finally, we plotted a posterior probability map of distribution in our sampling area.

Correlations environmental and genetics

To identify patterns of population genetic variation that derive from spatially-limited gene flow (i.e. isolation by distance, IBD), we conducted a Mantel test using a transformed genetic matrix (i.e. FST/(1-FST) and RST/(1-RST)), and geographic distance (i.e. logarithms of the linear distance between locations). Pearson correlation coefficients (i.e. r) were calculated in the VEGAN package of R functions [92], and p-values were calculated on 10,000 permutations. To identify average genetic diversity parameters (i.e. NA, HO, HE) that show correlations with average environmental variables (i.e. temperature, salinity, pH, oxygen, phosphate and nitrate), we conducted correlation analyses in the VEGAN package of R functions.

Environmental factors that promote changes at the microevolutionary level (i.e. population genetic structure) were estimated using hierarchical Bayesian models. We conducted analyses in GESTE v2 [93], in order to evaluate whether variables from our marine environmental dataset explain patterns of population genetics structure (specific factors and dataset used in GESTE were described above). Explicitly, GESTE relates FST values with environmental factors using a generalized linear model (GLM). We ran ten pilot runs (burn-in period) to have priors of mean and variance in the distribution of alpha parameters (alpha is the vector of regression coefficients that correspond to environmental data). After these pilot runs, we ran 10,000 MCMC iterations with a thinning interval of 100 and all those runs were replicated five times each. In all, combinations of marine environmental variables were considered and evaluated using estimates of posterior probability, and the degree of uncertainty of the estimations was measured by the 95% highest probability density interval (HPDI) [93]. In order to identify whether environmental variables could explain variations in allele frequencies among locations we conducted a redundancy analysis (RDA) in the VEGAN package of R functions. Specifically, we identified the relative contribution of each environmental variable on the allelic frequency variation using a forward stepwise selection (i.e. ordistep function) with the Akaike information criterion in VEGAN. P-values were estimate based on 10,000 permutations. The Pearson coefficient correlation (r) was estimated for only the environmental variables that better explain the data variability, in order to fulfill the non-correlation for multivariate analysis [94]. Finally, we plotted these environmental variables via the ordistep function.

Results

Overall, high genetic variability at all microsatellite loci were found for S. fuegensis samples, where the NA per locus ranged from 7 to 26, HE ranged from 0.636 to 0.955, and HO ranged from 0.148 to 0.957 (Table 1). The samples from Zone_D, located inside the TicToc Bay (East of the Corcovado Gulf, Fig 1) showed the lowest mean values of NA (12.8) and HE (0.803) (Table 1), and the surrounding Zone_E, located in the inner sea of Chiloé (North of the Corcovado Gulf, Fig 1), the highest values (see Table 1). Pairwise FST and RST indices showed a highly significant difference in comparison between Zone_D and the remainder locations (Table 2). The standardized measure of population differentiation F'ST and DST showed similar proportional magnitudes with respect to FST, and RST, where Zone D is the most divergent among sampling locations (S3 Table).

Significant deviations from the Hardy-Weinberg equilibrium were found at loci for some locations of samples due to homozygote excess as indicated by MICROCHECKER outcomes (Table 1). No pairwise comparison locus seems to be in linkage disequilibrium (P> 0.05).

Bayesian approaches based on the clustering method were congruent among them, despite the fact that STRUCTURE does not include a model that incorporates locus with possible null allele as in GENELAND. In addition, analyses that were run by each locus at a time showed a convergence in outcomes and found the same two clusters in 5 out of 8 loci (S4 Table). The smallest values of K that capture the major structure of the data were 2 in all of the cases (Fig 2): One genetic cluster (i.e. the major cluster, termed largest cluster, LC hereafter) includes almost all sampling localities (i.e. Zone_A, Zone_B, Zone_E, Zone_H, Zone_I, Zone_J, Zone_K, Zone_L, Zone_N); and another genetic cluster includes only Zone_D, a locality from the embayment area inside of the Tic-Toc MPA (East of the Corcovado Gulf; termed smallest cluster, SC hereafter). A similar outcome plot of membership was obtained even by using the admixture ancestry model (S1 Fig). The outcome plot from STRUCTURE revealed six individuals sampled from SC show a high genome proportion (>60%to multilocus genotype) from LC (Fig 2). Maps of posterior probabilities of population membership obtained from GENELAND to the SC and LC (Fig 2) showed the highest-probability lines (i.e. > 0.8 posterior probability), indicating the potential spatial position of genetic discontinuities between SC and LC at the mouth of Tic-Toc Bay. In addition, outcomes from GENELAND do not present any ghost populations at non-sampled areas, which means that all individuals where assigned to the number of populations inferred by the MCMC algorithm used by GENELAND.

thumbnail
Fig 2. Bayesian clustering results from STRUCTURE and GENELAND.

A) Plot shows the most likely number of clusters for the dataset. GENELAND analyses with posterior probability isoclines denoting the extent of genetic landscapes. Clusters indicated by GENELAND: B) Largest Cluster (LC) and C) Smallest Cluster (SC). Black dots represent localities analyzed in this study (represented by its respective letter) and regions with the greatest probability of inclusion are shown in white, whereas diminishing probabilities of inclusion are proportional to the degree of coloring.

https://doi.org/10.1371/journal.pone.0160670.g002

Transformed pairwise genetic distances between locations (FST/(1-FST) and RST/(1-RST)) and the natural log of geographic distance did not reveal any association of genetic distances with geography in the Mantel tests: low, non-significant negative correlations between distance matrices (i.e. FST (r = -0.1682, P = 0.7811) and RST (r = -0.1426, P = 0.6973)) were inferred, indicating the absence of isolation by distance, even when performing a posterior analysis and excluding the divergent zone D (i.e. FST (r = -0.3466, P = 0.9494) and RST (r = -0.0845, P = 0.5808)). The coefficient of determination (R2) between genetic diversity indices showed values rangingfrom 0.03 to 0.48. The relationship between HO and the average temperature was the highest correlation with a Pearson correlation coefficient (r) of 0.69.

The marine environmental factors that showed the highest sum of posterior probability included in the analyses were nitrate average and minimum; oxygen maximum; temperature maximum and range; and finally phosphate average, minimum and range (S5 Table). All these factors seem to be important in describing allelic frequency variation. Notwithstanding, none of the environmental variables showed a high sum of posterior probability, and the null model always explained more than 44% of the genetic structure among locations in each dataset (Table 3).

thumbnail
Table 3. Posterior probabilities of the three most probable models for the analyses including all the factors tested.

https://doi.org/10.1371/journal.pone.0160670.t003

Consequently, it means that no single environmental factor tested could explain the genetic structure observed in Bayesian analyses. The RDA that incorporated the significant environmental variables ordered by AIC (S6 Table) indicated that the minimum values of nitrate, and range values of temperature correlated withthe allelic frequency variation in our dataset (P < 0.039) (Fig 3). These explanatory variables were not correlated among them, where the variables showed a Pearson correlation coefficient of r = -0.113 among them (S7 Table) and determination coefficient of R = 0.012.

thumbnail
Fig 3. Redundancy analyses based on factors that show less Akaike value from ordistep analyses.

P-value was 0.039 (p<0.05). Open circles correspond to each zone, which are represented with its respective letter. Red crosses represent the allelic variability in the dataset, blue arrows point in the direction of maximum correlation, and the length of the arrow varies according to the strength of the correlation. RDA axis corresponds to an ordination constraint which represents a linear combination of these variables.

https://doi.org/10.1371/journal.pone.0160670.g003

Discussion

Our results show two genetic clusters in Sprattus fuegensis of the Patagonian fjords: one cluster restricted to Tic-Toc Bay (SC) and the other extending through the rest of the Chilean Patagonia (LC). We propose that the small cluster that is located at Tic-Toc MPA is the result of the singular oceanographic characteristics of this enclosed embayment. We find that temperature and nitrate correlate with the allelic frequency of the small cluster. We propose that larval retention and the effect of temperature and nitrate could be generating population genetic differentiation in this area. In addition, the spatial location of SC coincides with the location of the recently established marine protected area in Tic-Toc Bay, reinforcing the idea that this area has unique biogeographic characteristics.

Genetic diversity along Chilean Patagonia

The genetic variability of S. fuegensis is remarkably similar to the variability obtainedusing microsatellite in marine, anadromous and freshwater fishes by DeWoody and Avise [95]. Compared to other marine organisms, the heterozygosity (Mean HE = 0.80–0.90) in this study was similar to that of the congeneric S. sprattus (HE = 0.82–0.89 [61]) and it was higher than the values of their relatives Clupea pallasii (HE = 0.7–0.95, [96]) and C. harengus (HE = 0.71–0.78, [2], and HE 0.85–0.84, [97]).

The SC showed less allele number and less private alleles than LC (Table 1). Although this may be related to unequal sample size in each cluster, we discarded this possibility by conducting a rarefaction analysis, which showed no effects of sample size on allelic richness by location or by cluster. We propose that the low genetic variation showed by SC is promoted by larval retention, that generates low genetic flow and an enclosed small population size that is highly affected by genetic drift, which both change allele frequencies through time and therefore fixing alleles in this population. With the exception of private allele 306 at Spfu_29 locus from SC, all other private alleles showed low frequencies. In addition, some alleles from SC, despite being shared with the remaining locations, showed a 2- or 3-fold higher frequency than the remaining locations.

Genetic structure: biological characteristics and environmental features

Although none of the predicted clusters were found, two well-structured clusters were found by both Bayesian analyses, providing strong support for two genetic populations of S. fuegensis along its Chilean Patagonia distribution. The FST and RST indices were consistent with the Bayesian analyses, which means that our results are not dependent on the approach used. The result found in this study is in contrast with the genetic homogeneity found in other marine species in the same geographic area (Genypterus blacodes [98]). However, our FST and RST outcomes are quite comparable to fixation indices at neutral loci obtained for relatives to Patagonian sprat S. sprattus [61,99], Clupea harengus [2,100102] and Clupea pallasi [96,103,104]. For example, in Norwegian fjords, Glover et al. [61] found similar results in Sprattus sprattus, a closely related species of S. fuegensis. The authors found a small area that showed significant genetic differences between fjords versus the Southwest North Sea and the Southwest Celtic Sea. They further suggested that S. sprattus has a reduced connectivity between sea-going sprats and those found in Norwegian fjords. Nonetheless, they suggest that gene flow and demographic connectivity among the sprats inhabiting fjord locations is significant. In freshwater fishes distributed in Patagonia, it has been suggested that their genetic patterns are the result of barriers to gene flow and coastal refugees during glacial cycles (e.g. Percichthyidae [105], Galaxias maculatus [106]). In our case, given the current geographic distribution of S. fuegensis, and that it can tolerate a wide range of salinities, we cannot discard the hypothesis that historical refuges during the last glaciation might partly explain the observed pattern. Nonetheless, our data set based on microsatellite loci does not have the resolution to investigate this hypothesis, which should be evaluated using mitochondrial DNA.

The largest cluster (LC) found in our study extends from ~41° to ~53° LS (Fig 2). We hypothesize that the lack of genetic differentiation found in LC, in spite of the large geographic area covered, could be explained by the abundance and distribution of larvae, eggs and juveniles from nursery grounds or by the close proximity of spawning grounds along the Chilean geographic range. At present, specific spawning grounds of S. fuegensis have not been identified in its Chilean Patagonia distribution. Nonetheless, mature adults have been identified in the inner sea of Chiloé [37]. Moreover, the presence of juveniles has also been detected in the inner sea of Chiloé, and the fjord close to Aysén (i.e. between Puerto Aguirre and Estero Elefante) [50,107]. In numerous locations adjacent to the Strait of Magellan, Pacific and Atlantic Oceans on the Magellanic shelf, the presence of eggs has been discovered[13,42,44]. Using otolith microchemistry from juvenile S. fuegensis individuals, Galleguillos et al. [108] showed the presence of three different nursery grounds, which can be found in the inner sea of Chiloé, the fjord close to Aysén and in the Strait of Magellan. In the area of the Strait of Magellan and channels adjacent to the Atlantic Ocean, Sánchez et al. [42] identified the largest nursery ground of S. fuegensis along the Argentinean Patagonian coast with a juvenile production of 1.3x109 individuals. Similarly, a probable explanation of the non-genetic differences found in Genypterus blacodes along inner waters, channels and fjords in Chilean Patagonia, was the close proximity of spawning grounds in the same study area [98]. Adult migration in S. fuegensis has not been recorded to date, however, indirect evidence (i.e. microchemistry of otoliths and parasites tags) has pointed out that an active dispersal of adults must exist between the inner sea of Chiloé and the fjord close to Aysén [108]. The same mechanism has been proposed in Engraulis ringens [109] and Strangomera bentinki [110], two small pelagic marine fishes distributed along the continental shelf. Overall, taking into account the broad distribution of eggs, larvae and juveniles that has been recorded for this species [13,42,44,50,107], we can suggest that migration via passive dispersal might be playing a key role in the lack of genetic structure found within LC.

The environmental characteristics can also explain the low genetic differentiation of LC. Water bodies and circulation patterns could be causing migration via passive dispersal [56]. Sievers and Silva [15] recorded the directionality of different bodies of water along the Patagonian Chilean sea (S2 Fig). They described in the superficial level (i.e. 0- ~30 m) a narrow estuarine water layer with low salinity that leads into the Boca del Guafo [15]. At the middle level (i.e. ~30- ~150 m), a depth where mainly S. fuegensis can be recorded, they described a broad Subantarctic body of water that goes into Boca del Guafo and then divides northward intot he inner sea of Chiloé and southward to the fjord and channels close to Aysén [15]. Therefore, there is a superficial circulation pattern through all the extend of LC that would be driving the connectivity among localities.

The smallest cluster (SC) has a restricted geographic distribution inside Tic-Toc Bay. The SC showed highly significant differences, giving strong support for its existence. Based on GESTE analyses, none of the tested environmental variables, physical (i.e. temperature) or chemical (salinity, pH, oxygen, phosphate and nitrate) of the datasets incorporated in this study were better than the null model (Table 3). The RDA showed similar results. However, minimum nitrate and range temperature were variables that explain the allelic frequency variation in the two clusters found in this study (Fig 3). We found contradictory outcomes in landscape genetics analysis using the GESTE and the RDA approaches. For Bayesian analysis conducted in GESTE, we did not found any variable(s) that fit better than the null model, which would indicate that environmental data do not have any effect on the genetic structure observed. However, Foll and Gaggiotti[93] indicated that when GESTE fails to identify the true model, the outcomes only are not conclusive. For multivariate analysis conducted in RDA, two variables showed a significant contribution of the genetic structure in Zone_D. The RDA analysis has been strongly supported as a powerful approach in landscape genetics as noted by Legendre and Fortin [94]. These kinds of contradictions were observed by Balkenhol et al. [111], when they comparing eleven methods commonly used to link landscape and genetics data which indicated that nonlinear methods in multivariate analysis have a better success rate (i.e. in our case RDA) than others, including GESTE. However, this does not mean that GESTE is unsuitable for landscape genetics analysis, but this analysis should be performed together with other approaches in order to choose optimal combinations of landscape genetics methods [111]. Therefore, we put more emphasis on RDA outcomes than GESTE outcomes in this study.

The SC was an unexpected outcome considering that, based on previous environmental information whereby we expected to find genetic differences between the more isolated areas. The SC is localized within the Chiloense Marine Ecoregion [112], an ecoregion that has been described as having an upwelling system where mesoscale processes such as eddies, fronts and plumes increase the retention of phytoplankton [113] and produce highly productive spring and autumn seasons [114,115]. A recent study showed that features such as eddies and fronts can enhance and concentrate the marine productivity which promote the generation of high quality patches in the plankton to be used by pelagic larvae, enhancing their survival [116]. Therefore, high phytoplankton and zooplankton aggregations and kelp forests provide feeding and refuge to diverse fish and invertebrate communities [113], and produce an overall pattern of high biodiversity [117,118]. Davila et al. [119] propos that the entire area functions as a large estuarine system. Accordingly, in the area of the Gulf of Corcovado-Boca del Guafo several submarine topographic features, groups of islands and coastal narrowing, determine a geographical configuration that energize and differentiate the enclosing water bodies [113], and that may be promoting the isolation of Tic-Toc Bay where the SC is located. Actually, the surface layer (0–~30m) does not enter Tic-Toc Bay (S2 Fig), which would decrease even more the gene flow with LC by decreasing the transport of pelagic larvae of S. fuegensis from outside the bay. Nonetheless, we found individuals from LC within SC and vice versa, and we propose that they are the result of adult migration between these areas, preferentially from LC to SC.

This SC is concordant with the marine protected area (MPA) created in 2014 by the Chilean government, which was based on the high biodiversity and unique biotic and abiotic features of the zone. The Tic-Toc MPA has a surface area of 97,929 ha surrounding the Corcovado Gulf (Fig 1) and according to Alvarez et al. [113] it has a high diversity and abundance of cetaceans, dolphins, and other marine mammals and it serves as a refuge for several taxa (i.e. phytoplankton, zooplankton, kelp forests, fish and invertebrate communities [113]). In addition, this zone is one of the few fjord areas in Chile where there is no aquaculture. In this area, so far, only species and ecosystem diversity have been considered. Our results provide the first evidence of the importance that this zone could have in regard to intraspecific genetic diversity, supporting even further its uniqueness and justifying its protection. In future similar studies it would be interesting to incorporate other marine organisms that show comparable and contrasting life history traits in order to investigate how the oceanographic features of this area could be determining their uniqueness. Our study also reminds us of the importance of incorporating genetic diversity in the analyses of future conservation areas whenever this information is available and to not underestimate the contribution to the preservation of biodiversity that a particular zone could providing.

In conclusion, our data show that the singular genetic differences found inside the Tic-Toc MPA are the result of genetic drift, probably due to larval retention throughout a combination of oceanographic mesoscale processes, geographical configuration, and the local effect of the environmental variables on genetic variation. These features have generated the isolated and restricted area that promoted genetic differentiation. Further analyses should be carried out to confirm this spatial genetic pattern, test whether this pattern is stable in the long term and also whether environmental features not explicitly tested in the present study (i.e. currents of water bodies) are able to better explain the population genetic structure of this species.

Supporting Information

S1 Table. Environmental variables used in our analyses.

Ave: average, Rang: Range, Max: Maximum, Min: Minimum.

https://doi.org/10.1371/journal.pone.0160670.s001

(DOCX)

S2 Table. Bayes factor comparison among different models incorporated in STRUCTURE and GENELAND software.

AdUn: admixture and uncorrelated model, AdCo: admixture and correlated model, NAdUn: no admixture and uncorrelated model, NAdCo: no admixture and correlated model.

https://doi.org/10.1371/journal.pone.0160670.s002

(DOCX)

S3 Table. The standardized measure of population differentiation F'ST under the diagonal and DST above the diagonal.

https://doi.org/10.1371/journal.pone.0160670.s003

(DOCX)

S4 Table. Cluster number by GENELAND assigned to each locus based on posterior probability density and assignment of each location to cluster found per locus.

https://doi.org/10.1371/journal.pone.0160670.s004

(DOCX)

S5 Table. Sum of posterior probabilities of models that include a given factor.

GESTE analyses included all 6 factors. Bold value indicates the two highest factor scores.

https://doi.org/10.1371/journal.pone.0160670.s005

(DOCX)

S6 Table. Relative contribution of each environmental variable tested using Akaike’s information criterion.

Phos: Phosphate, Lat: Latitude, Long: Longitude, Nit: Nitrate, Oxy: Oxygen, Sal: Salinity, Tem: Temperature. Ave: average, Rang: Range, Max: Maximum, Min: Minimum. Bold values show significant p-values. Variable kept means environmental variables that explain variation in allele frequencies among locations.

https://doi.org/10.1371/journal.pone.0160670.s006

(DOCX)

S7 Table. Pearson coefficient per each environmental variable tested.

Bold values show the correlation coefficient between environmental variables kept in the RDA analysis.

https://doi.org/10.1371/journal.pone.0160670.s007

(DOCX)

S1 Fig. Plots to estimate the best number of genetic clusters.

A) Evanno et al. [88] plot for detecting the number of K groups that best fit the data. B) Plot of the mean likelihood L(K) and variance per K value from STRUCTURE. C) Plot of the number of populations simulated from the posterior distribution obtained with GENELAND.

https://doi.org/10.1371/journal.pone.0160670.s008

(EPS)

S2 Fig. Schematic map of horizontal water circulation in different depth layers, including sample locations.

A) surface layer (0–~30m); B) intermediate layer (~30–~150m); and C) deep layer (~150 m to bottom of the sea). Image modified from Sievers and Silva [15]. Sample locations in black dots.

https://doi.org/10.1371/journal.pone.0160670.s009

(EPS)

S1 Dataset. Dataset of loci microsatellites used in this study.

https://doi.org/10.1371/journal.pone.0160670.s010

(XLSX)

Acknowledgments

The authors are very grateful to Fernanda X. Oyarzun for comments and suggestions, which greatly improved the final version of the manuscript. We thank the editor and two anonymous reviewers for their constructive comments, which helped us to improve the manuscript. This manuscript was one the chapters of the PhD thesis of CCA. CCA and SFF were supported by Doctoral Fellowships for the ‘Programa de Doctorado en Sistemática y Biodiversidad’, from the graduate school of the Universidad de Concepción, Chile. SFF was supported by a CONICYT doctoral fellowship. CEH was supported by FONDECYT. CCA, SFF, RG, CEH were supported by Fondo de Investigación Pesquera (FIP 2010–17).

Author Contributions

  1. Conceptualization: CCA SFF RG CEH.
  2. Data curation: CCA.
  3. Formal analysis: CCA.
  4. Funding acquisition: CCA SFF RG.
  5. Investigation: CCA SFF.
  6. Methodology: CCA SFF RG.
  7. Project administration: SFF RG.
  8. Resources: CCA SFF RG CEH.
  9. Software: CCA.
  10. Supervision: CCA SFF RG CEH.
  11. Validation: CCA SFF.
  12. Visualization: CCA.
  13. Writing - original draft: CCA CEH.
  14. Writing - review & editing: CCA CEH.

References

  1. 1. Selkoe KA, Henzler CM, Gaines SD. Seascape genetics and the spatial ecology of marine populations. Fish Fish. 2008;9: 363–377.
  2. 2. Jørgensen HBH, Hansen MM, Bekkevold D, Ruzzante DE, Loeschcke V. Marine landscapes and population genetic structure of herring (Clupea harengus L.) in the Baltic Sea. Mol Ecol. Blackwell Science Ltd; 2005;14: 3219–3234.
  3. 3. Galindo HM, Olson DB, Palumbi SR. Seascape genetics: a coupled oceanographic-genetic model predicts population structure of Caribbean corals. Curr Biol. 2006;16: 1622–6. pmid:16920623
  4. 4. Manel S, Schwartz MK, Luikart G, Taberlet P. Landscape genetics: combining landscape ecology and population genetics. Trends Ecol Evol. 2003;18: 189–197.
  5. 5. Storfer A, Murphy MA, Evans JS, Goldberg CS, Robinson S, Spear SF, et al. Putting the “landscape” in landscape genetics. Heredity (Edinb). 2007;98: 128–42.
  6. 6. Storfer A, Murphy MA, Spear SF, Holderegger R, Waits LP. Landscape genetics: where are we now? Mol Ecol. 2010;19: 3496–514. pmid:20723061
  7. 7. Riginos C, Liggins L. Seascape Genetics: Populations, Individuals, and Genes Marooned and Adrift. Geogr Compass. 2013;7: 197–216.
  8. 8. Johnson DW. The nature and origin of fjords. Science (80-). 1915;41: 537–543.
  9. 9. Holtedahl H. Notes on the Formation of Fjords and Fjord-Valleys. Geogr Ann Ser A, Phys Geogr. 1967;49: 188–203.
  10. 10. Syvitski JPM, Shaw J. Sedimentology and Geomorphology of Fjords. Geomorphol Sedimentol Estuaries. 1995;53: 113–178.
  11. 11. Syvitski JP, Burrell DC, Skei JM. Fjords and Their Study. Fjords. Springer New York; 1987. pp. 3–17.
  12. 12. Pantoja S, Luis Iriarte J, Daneri G. Oceanography of the Chilean Patagonia. Cont Shelf Res. 2011;31: 149–153.
  13. 13. Landaeta MF, López G, Suárez-Donoso N, Bustos CA, Balbontín F. Larval fish distribution, growth and feeding in Patagonian fjords: potential effects of freshwater discharge. Environ Biol Fishes. 2012;93: 73–87.
  14. 14. Sievers HA. Temperatura y salinidad en canales y fiordos australes. In: Silva N, Palm S, editors. Avances en el conocimiento oceanográfico de las aguas interiores chilenas, Puerto Montt a Cabo de Hornos. Valparaíso, Chile.; 2006. pp. 31–36.
  15. 15. Sievers HA, Silva N. Masas de agua y circulación en los canales y fiordos australes. In: Silva N, Palma S, editors. Avances en el conocimiento oceanográfico de las aguas interiores chilenas, Puerto Montt a Cabo de Hornos. Valparaíso, Chile.: Comité Oceanográfico Nacional, Pontificia Universidad Católica de Valparaíso; 2006. pp. 53–58.
  16. 16. Taylor MS, Hellberg ME. Genetic Evidence for Local Retention of Pelagic Larvae in a Caribbean Reef Fish. Science (80-). 2003;299: 107–109.
  17. 17. Coleman MA, Feng M, Roughan M, Cetina-Heredia P, Connell SD. Temperate shelf water dispersal by Australian boundary currents: Implications for population connectivity. Limnol Oceanogr Fluids Environ. Duke University Press; 2013;3: 295–309.
  18. 18. Pespeni MH, Chan F, Menge BA, Palumbi SR. Signs of adaptation to local pH conditions across an environmental mosaic in the California Current Ecosystem. Integr Comp Biol. 2013;53: 857–70. pmid:23980118
  19. 19. Reusch TBH. Climate change in the oceans: evolutionary versus phenotypically plastic responses of marine animals and plants. Evol Appl. 2014;7: 104–22. pmid:24454551
  20. 20. Rijnsdorp AD, Peck MA, Engelhard GH, Mollmann C, Pinnegar JK. Resolving the effect of climate change on fish populations. ICES J Mar Sci. 2009;66: 1570–1583.
  21. 21. Basedow SL, Eiane K, Tverberg V, Spindler M. Advection of zooplankton in an Arctic fjord (Kongsfjorden, Svalbard). Estuar Coast Shelf Sci. 2004;60: 113–124.
  22. 22. Willis K, Cottier F, Kwasniewski S, Wold A, Falk-Petersen S. The influence of advection on zooplankton community composition in an Arctic fjord (Kongsfjorden, Svalbard). J Mar Syst. 2006;61: 39–54.
  23. 23. Beuchel F, Gulliksen B, Carroll ML. Long-term patterns of rocky bottom macrobenthic community structure in an Arctic fjord (Kongsfjorden, Svalbard) in relation to climate variability (1980–2003). J Mar Syst. 2006;63: 35–48.
  24. 24. Kristoffersen JB, Salvanes AG V. Life history of Maurolicus muelleri in fjordic and oceanic environments. J Fish Biol. 1998;53: 1324–1341.
  25. 25. Kristoffersen JB, Salvanes AGV. Distribution, growth, and population genetics of the glacier lanternfish (Benthosema glaciale) in Norwegian waters: Contrasting patterns in fjords and the ocean. Mar Biol Res. Taylor & Francis; 2009;5: 596–604.
  26. 26. Castellani M, Rosland R, Urtizberea A, Fiksen Ø. A mass-balanced pelagic ecosystem model with size-structured behaviourally adaptive zooplankton and fish. Ecol Modell. 2013;251: 54–63.
  27. 27. Goodson MS, Giske J, Rosland R. Growth and ovarian development of Maurolicus muelleri during spring. Mar Biol. 1995;124: 185–195.
  28. 28. Olsen RB, Richardson K, Simonsen V. Population differentiation of eelpout Zoarces viviparus in a Danish fjord. Mar Ecol Prog Ser. 2002;227: 97–107. Available: http://www.int-res.com/abstracts/meps/v227/p97-107/
  29. 29. Kaartvedt S, Røstad A, Klevjer TA. Sprat Sprattus sprattus can exploit low oxygen waters for overwintering. Mar Ecol Prog Ser. 2009;390: 237–249. Available: http://www.int-res.com/abstracts/meps/v390/p237-249/
  30. 30. Balbontín F. Ictioplancton de los canales y fiordos australes. In: Silva N, Palma S, editors. Avances en el conocimiento oceanográfico de las aguas interiores chilenas, Puerto Montt a cabo de Hornos. Valparaíso, Chile.: Comité Oceanográfico Nacional, Pontificia Universidad Católica de Valparaíso; 2006. pp. 115–120.
  31. 31. Bustos CA, Balbontín F, Landaeta MF. Spawning of the southern hake Merluccius australis (Pisces: Merlucciidae) in Chilean fjords. Fish Res. 2007;83: 23–32.
  32. 32. Bradbury IR, Snelgrove PVR, Bentzen P, de Young B, Gregory RS, Morris CJ. Structural and functional connectivity of marine fishes within a semi-enclosed Newfoundland fjord. J Fish Biol. 2009;75: 1393–409. pmid:20738621
  33. 33. Fevolden SE, Westgaard JI, Pedersen T, Præbel K. Settling-depth vs. genotype and size vs. genotype correlations at the Pan I locus in 0-group Atlantic cod Gadus morhua. Mar Ecol Prog Ser. 2012;468: 267–278. Available: http://www.int-res.com/abstracts/meps/v468/p267-278/
  34. 34. Cousseau MB. Revisión taxonómica y análisis de los caracteres morfométricos y:merísticos de la sardina f ueguina, Sprattus fuegensis (Jenyns, 1842) (Pisces, Clupeidae). Rev Investig y Desarro Pesq. 1982;3: 77–94. Available: http://oceandocs.net/handle/1834/2054
  35. 35. Whitehead PJP. FAO species catalogue. Vol.7. Clupeoid fishes of the world. An annotated and illustrated catalogue of the herrings, sardines, pilchards, sprats, anchovies and wolf- herrings. Part 1—Chirocentridae, Clupeidae and Pristigasteridae. FAO Fish Synopsis. FAO Fisher. FAO species catalogue.; 1985;125: 1–303.
  36. 36. Pequeño G. Peces de Chile. Lista sistemática revisada y comentada. Rev Biol Mar Valparaíso. 1989;24: 1–132.
  37. 37. Aranis A, Meléndez R, Germán P, Cerna T F. Sprattus fuegensis en aguas interiores de Chiloé, Chile (Osteichthyes: Clupeiformes:Clupeidae). Gayana. 2007;71: 102–113.
  38. 38. Leal E, Canales TM, Aranis A, Gonzáles M. Actividad reproductiva y longitud de madurez de sardina austral Sprattus fuegensis en el mar interior de Chiloé, Chile. Rev Biol Mar Oceanogr. 2011;46: 43–51.
  39. 39. Cerna F, Leal E, López A, Plaza G. Age, growth and natural mortality of the Patagonian sprat Sprattus fuegensis (Jenyns, 1842) in Chiloé inland sea, southern Chile. Lat Am J Aquat Res. 2014;42: 580–587.
  40. 40. Shirakova EN. Some biological features of Tierra del Fuego sprat. Sov J Mar Biol. 1978;4: 697–702.
  41. 41. Shirakova EN. Contribution to the biology of Tierra del Fuego sprat Sprattus fuegensis (Jenyns, 1842). Biol Sea. 1978;3: 78–84.
  42. 42. Sánchez RP, Remeslo A, Madirolas A, de Ciechomski JD. Distribution and abundance of post-larvae and juveniles of the Patagonian sprat, Sprattus fuegensis, and related hydrographic conditions. Fish Res. 1995;23: 47–81.
  43. 43. Bustos CA, Landaeta MF, Balbontín F. Efectos ambientales sobre la variabilidad espacial del ictioplancton de Chile austral durante noviembre de 2005. Rev Chil Hist Nat. 2008;81: 205–219. Available: http://www.scielo.cl/scielo.php?pid=S0716-078X2008000200005&script=sci_arttext&tlng=e
  44. 44. Landaeta MF, Bustos CA, Palacios-Fuentes P, Rojas P, Balbontín F. Distribucion del ictioplancton en la Patagonia austral de Chile: potenciales efectos del deshielo de Campos de Hielo Sur. Lat Am J Aquat Res. 2011;39: 236–249.
  45. 45. Hansen JE. Estimación de parámetros poblacionales del efectivo de sardina fueguina (Sprattus fuegensis) de la costa continental Argentina [Internet]. INIDEP Informe Técnico 27, Mar del Plata, Argentina; 1999. Available: http://www.oceandocs.org/handle/1834/2530
  46. 46. Sievers HA, Calvete C, Silva N. Distribución de características físicas, masas de agua y circulación general para algunos canales australes entre el Golfo de Penas y el Estrecho de Magallanes (Crucero CIMAR-FIORDO 2), Chile. Cienc y Tecnol del Mar. 2002;25: 17–43.
  47. 47. Silva N, Calvete C. Características oceanográficas físicas y químicas de canales australes chilenos entre el Golfo de Penas y el Estrecho de Magallanes. Cienc y Tecnol del Mar. 2002;25: 23–88.
  48. 48. Valdenegro A, Silva N. Caracterización oceanográfica física y química de la zona de canales y fiordos australes de Chile entre Estrecho de Magallanes y Cabo de Hornos (CIMAR 3 Fiordos). Cienc y Tecnol del Mar. 2003;26: 5–44.
  49. 49. Subpesca. Estado de situación de las principales pesquerías chilenas al año 2014. 2015.
  50. 50. Niklitschek E, Toledo P, Hernaández E, Nelson J, Soule M, Herranz C, et al. Evaluación hidroacústica de pequeños pelágicos en aguas interiores de la X y XI regiones, año 2007. Final Report FIP N° 2007–05; 2009.
  51. 51. Carrasco C, Silva N. Comparación de las características oceanográficas físicas y químicas presentes en la zona de Puerto Montt a la Boca del Guafo entre el invierno y la primavera de 2004 y entre las primaveras de 1995 y 2004. Cienc y Tecnol del Mar. 2010;33: 17–44.
  52. 52. Guzmán D, Silva N. Caracterización física y química y masa de agua en los canales australes de Chile entre Boca del Guafo y Golfo Elefantes (Crucero CIMAR-Fiordo 4). Cienc y Tecnol del Mar. 2002;25: 45–76.
  53. 53. Silva N, Guzmán D. Condiciones oceanográficas físicas y químicas, entre la Boca del Guafo y Fiordo Aysén (Crucero CIMAR 7 Fiordos). Cienc y Tecnol del Mar. 2006;29: 25–44.
  54. 54. Silva N, Valdenegro A. Caracterización oceanográfica de canales australes chilenos entre la Boca del Guafo y los canales Pulluche—Chacabuco (CIMAR 8 fiordos)*. Cienc y Tecnol del Mar. 2008;31: 5–44.
  55. 55. Silva N, Calvete C, Sievers HA. Características oceanográficas físicas y químicas de canales australes chilenos entre Puerto Montt y Laguna San Rafael (Crucero Cimar-Fiordo 1). Cienc y Tecnol del Mar. 1997;20: 23–106.
  56. 56. Silva N, Calvete C, Sievers HA. Masas de agua y circulación general para algunos canales australes entre Puerto Montt y Laguna San Rafael, Chile (Crucero Cimar-Fiordo 1). Cienc y Tecnol del Mar. 1998;21: 17–48.
  57. 57. Sievers HA. Temperature and salinity in the austral Chilean channels and fjords. In: Silva N, Palma S, editors. Progress in the oceanographic knowledge of Chilean interior waters, from Puerto Montt to Cape Horn. Valparaíso, Chile.: Comité Oceanográfico Nacional, Pontificia Universidad Católica de Valparaíso; 2006. pp. 31–36.
  58. 58. Silva N. Dissolved oxygen, pH, and nutrients in the austral Chilean channels and fjords. In: Silva N, Palma S, editors. Progress in the oceanographic knowledge of Chilean interior waters, from Puerto Montt to Cape Horn. Valparaíso, Chile.: Comité Oceanográfico Nacional, Pontificia Universidad Católica de Valparaíso; 2006. pp. 37–43.
  59. 59. Waples RS. Separating the wheat from the chaff: patterns of genetic differentiation in high gene flow species. J Hered. 1998;89: 438–450.
  60. 60. Allendorf FW, Phelps SR. Use of Allelic Frequencies to Describe Population Structure. Can J Fish Aquat Sci. 1981;38: 1507–1514.
  61. 61. Glover KA, Skaala Ø, Limborg M, Kvamme C, Torstensen E. Microsatellite DNA reveals population genetic differentiation among sprat (Sprattus sprattus) sampled throughout the Northeast Atlantic, including Norwegian fjords. ICES J Mar Sci. 2011;68: 2145–2151.
  62. 62. Valdenegro A, Silva N. Caracterización oceanográfica física y química de la zona de canales y fiordos australes de Chile entre el Estrecho de Magallanes y Cabo de Hornos (CIMAR 3 FIORDOS). Cienc y Tecnol del Mar. 2003;26: 19–60.
  63. 63. Ferrada-Fuentes S, Galleguillos R, Canales-Aguirre CB, Love CN, Jones KL, Lance SL. Development and characterization of thirty-three microsatellite markers for the Patagonian sprat, Sprattus fuegensis (Jenyns, 1842), using paired-end Illumina shotgun sequencing. Conserv Genet Resour. 2014;
  64. 64. Kalinowski ST. Counting Alleles with Rarefaction: Private Alleles and Hierarchical Sampling Designs. Conserv Genet. 2004;5: 539–543.
  65. 65. Kalinowski ST. HP-RARE 1.0: a computer program for performing rarefaction on measures of allelic richness. Mol Ecol Notes. 2005;5: 187–189.
  66. 66. van Oosterhout C, Hutchinson WF, Wills DPM, Shipley P. MICRO-CHECKER: software for identifying and correcting genotyping errors in microsatellite data. Mol Ecol Notes. Blackwell Science Ltd; 2004;4: 535–538.
  67. 67. Chapuis M-P, Estoup A. Microsatellite null alleles and estimation of population differentiation. Mol Biol Evol. 2007;24: 621–631. pmid:17150975
  68. 68. Girard P, Angers B. Assessment of power and accuracy of methods for detection and frequency-estimation of null alleles. Genetica. 2008;134: 187–97. pmid:18060508
  69. 69. Carlsson J. Effects of microsatellite null alleles on assignment testing. J Hered. 2008;99: 616–623. pmid:18535000
  70. 70. Guillot G, Santos F, Estoup A. Analysing georeferenced population genetics data with Geneland: a new algorithm to deal with null alleles and a friendly graphical user interface. Bioinforma. 2008;24: 1406–1407.
  71. 71. Corander J, Waldmann P, Sillanpaa MJ. Bayesian analysis of genetic differentiation between populations. Genetics. 2003;163: 367–374. Available: http://www.genetics.org/content/163/1/367.abstract pmid:12586722
  72. 72. Chapuis M-P, Lecoq M, Michalakis Y, Loiseau A, Sword GA, Piry S, et al. Do outbreaks affect genetic population structure? A worldwide survey in Locusta migratoria, a pest plagued by microsatellite null alleles. Mol Ecol. 2008;17: 3640–53. pmid:18643881
  73. 73. Guinand B, Scribner KT, Page KS, Filcek K, Main L, Burnham-Curtis MK. Effects of coancestry on accuracy of individual assignments to population of origin: examples using Great Lakes lake trout (Salvelinus namaycush). Genetica. 2006;127: 329–340. pmid:16850237
  74. 74. Anderson EC, Dunham KK. The influence of family groups on inferences made with the program Structure. Mol Ecol Resour. Blackwell Publishing Ltd; 2008;8: 1219–1229.
  75. 75. Rodríguez-Ramilo ST, Wang J. The effect of close relatives on unsupervised Bayesian clustering algorithms in population genetic structure analysis. Mol Ecol Resour. Blackwell Publishing Ltd; 2012;12: 873–884.
  76. 76. Jones OR, Wang J. COLONY: a program for parentage and sibship inference from multilocus genotype data. Mol Ecol Resour. Blackwell Publishing Ltd; 2010;10: 551–555.
  77. 77. Wang J, Santure AW. Parentage and Sibship Inference From Multilocus Genotype Data Under Polygamy. Genetics. 2009;181: 1579–1594. pmid:19221199
  78. 78. Wang J. Sibship Reconstruction From Genetic Data With Typing Errors. Genetics. 2004;166: 1963–1979. pmid:15126412
  79. 79. Peakall R, Smouse PE. GenAlEx 6.5: genetic analysis in Excel. Population genetic software for teaching and research—an update. Bioinformatics. 2012;28: 2537–9. pmid:22820204
  80. 80. Excoffier L, Lischer HEL. Arlequin suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Mol Ecol Resour. 2010;10: 564–7. pmid:21565059
  81. 81. Raymond M, Rousset F. GENEPOP (Version 1.2): Population Genetics Software for Exact Tests and Ecumenicism. J Hered. 1995;86: 248–249. Available: http://jhered.oxfordjournals.org/content/86/3/248.short
  82. 82. Rousset F. Genepop’007: a complete re-implementation of the genepop software for Windows and Linux. Mol Ecol Resour. 2008;8: 103–106. pmid:21585727
  83. 83. Meirmans PG, Hedrick PW. Assessing population structure: FST and related measures. Mol Ecol Resour. Blackwell Publishing Ltd; 2011;11: 5–18.
  84. 84. Rice WR. Analyzing tables of statistical test. Evolution (N Y). 1989;43: 223–225. Available: http://www.jstor.org/stable/2409177
  85. 85. Pritchard JK, Stephens M, Donnelly P. Inference of population structure using multilocus genotype data. Genetics. 2000;155: 945–959. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1461096&tool=pmcentrez&rendertype=abstract pmid:10835412
  86. 86. Falush D, Stephens M, Pritchard JK. Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics. 2003;164: 1567–1587. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1462648&tool=pmcentrez&rendertype=abstract pmid:12930761
  87. 87. Hubisz MJ, Falush D, Stephens M, Pritchard JK. Inferring weak population structure with the assistance of sample group information. Mol Ecol Resour. 2009;9: 1322–32. pmid:21564903
  88. 88. Evanno G, Regnaut S, Goudet J. Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol Ecol. 2005;14: 2611–20. pmid:15969739
  89. 89. Earl DA, VonHoldt BM. STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conserv Genet Resour. Springer Netherlands; 2012;4: 359–361.
  90. 90. Guillot G, Mortier F, Estoup A. Geneland: a computer package for landscape genetics. Mol Ecol Notes. Blackwell Science Ltd; 2005;5: 712–715.
  91. 91. Guillot G, Estoup A, Mortier F, Cosson JF. A spatial statistical model for landscape genetics. Genetics. 2005;170: 1261–1280. pmid:15520263
  92. 92. Oksanen J, Blanchet, F. Guillaume Kindt R, Legendre P, Minchin PR, O’Hara RB, Simpson GL, et al. VEGAN: community ecology package. R package version 2.0–10. http://CRAN.R-project.or. 2013.
  93. 93. Foll M, Gaggiotti O. Identifying the environmental factors that determine the genetic structure of populations. Genetics. 2006;174: 875–891. pmid:16951078
  94. 94. Legendre P, Fortin M-J. Comparison of the Mantel test and alternative approaches for detecting complex multivariate relationships in the spatial analysis of genetic data. Mol Ecol Resour. Blackwell Publishing Ltd; 2010;10: 831–844.
  95. 95. DeWoody JA, Avise JC. Microsatellite variation in marine, freshwater and anadromous fishes compared with other animals. J Fish Biol. 2000;56: 461–473.
  96. 96. Semenova A V., Stroganov AN, Smirnov AA, Afanas’ev KI, Rubtsova GA. Genetic variations in Clupea pallasii herring from Sea of Okhotsk based on microsatellite markers. Russ J Genet. 2014;50: 175–179.
  97. 97. Mariani S, Hutchinson WF, Hatfield EMC, Ruzzante DE, Simmonds EJ, Dahlgren TG, et al. North Sea herring population structure revealed by microsatellite analysis. Mar Ecol Prog Ser. 2005;303: 245–257. Available: http://www.int-res.com/abstracts/meps/v303/p245-257/
  98. 98. Canales-Aguirre CB, Ferrada S, Hernández CE, Galleguillos R. Population structure and demographic history of Genypterus blacodes using microsatellite loci. Fish Res. Elsevier B.V.; 2010;106: 102–106.
  99. 99. Limborg MT, Pedersen JS, Hemmer-Hansen J, Tomkiewicz J, Bekkevold D. Genetic population structure of European sprat Sprattus sprattus: differentiation across a steep environmental gradient in a small pelagic fish. Mar Ecol Prog Ser. 2009;379: 213–224.
  100. 100. Jørgensen HBH, Pertoldi C, Hansen MM, Ruzzante DE, Loeschcke V. Genetic and environmental correlates of morphological variation in a marine fish: the case of Baltic Sea herring (Clupea harengus). Can J Fish Aquat Sci. NRC Research Press Ottawa, Canada; 2008;65: 389–400.
  101. 101. Shaw PW, Turan C, Wright JM, O’Connell M, Carvalho GR. Microsatellite DNA analysis of population structure in Atlantic herring (Clupea harengus), with direct comparison to allozyme and mtDNA RFLP analyses. Heredity (Edinb). 1999;83: 490–499.
  102. 102. Teacher AG, André C, Jonsson PR, Merilä J. Oceanographic connectivity and environmental correlates of genetic structuring in Atlantic herring in the Baltic Sea. Evol Appl. 2013;6: 549–67. pmid:23745145
  103. 103. Sugaya T, Sato M, Yokoyama E, Nemoto Y, Fujita T, Okouchi H, et al. Population genetic structure and variability of Pacific herring Clupea pallasii in the stocking area along the Pacific coast of northern Japan. Fish Sci. 2008;74: 579–588.
  104. 104. Wildes SL, Vollenweider Johanna J. Nguyen HT, Guyon JR. Genetic variation between outer-coastal and fjord populations of Pacific herring (Clupea pallasii) in the eastern Gulf of Alaska. Fish Bull. 2010;109: 382–393.
  105. 105. Ruzzante DE, Walde SJ, Cussac VE, Dalebout ML, Seibert J, Ortubay S, et al. Phylogeography of the Percichthyidae (Pisces) in Patagonia: roles of orogeny, glaciation, and volcanism. Mol Ecol. 2006;15: 2949–2968. pmid:16911213
  106. 106. Zemlak TS, Habit EM, Walde SJ, Carrea C, Ruzzante DE. Surviving historical Patagonian landscapes and climate: molecular insights from Galaxias maculatus. BMC Evol Biol. 2010;10: 67. pmid:20211014
  107. 107. Cárdenas G. Análisis de la composición de tallas, indicadores nutricionales y parámetros de crecimiento Strangomera bentincki (sardina común) y Sprattus fuegensis (sardina austral) en el sur de Chile. Tesis para optar al grado de Biología Marina, Universidad Austral de Chile. 2009.
  108. 108. Galleguillos R, Ferrada-Fuentes S, Canales-Aguirre CB, Hernández CE, Oliva ME, González MT, et al. Determinación de unidades poblacionales de sardina austral entre la X y XII regiones de Chile. [Internet]. Final Report FIP N°2010–17; 2012. Available: www.fip.cl
  109. 109. Ferrada S, Hernández K, Montoya R, Galleguillos R. Estudio poblacional del recurso anchoveta (Engraulis ringens Jenyns 1842) (Clupeiformes, Engraulidae), mediante análisis de ADN. Gayana. 2002;66: 243–248. Available: http://www.scielo.cl/scielo.php?pid=S0717-65382002000200022&script=sci_arttext
  110. 110. Galleguillos R, Troncoso L, Monsalves J, Oyarzun C. Diferenciación poblacional en la sardina chilena Strangomera bentincki (Pisces: Clupeidae) análisis genético de variabilidad proteínica. Rev Chil Hist Nat. 1997;70: 351–361. Available: http://www2.udec.cl/~acuigen/upload/file/Galleguillosetal_1997.pdf
  111. 111. Balkenhol N, Waits LP, Dezzani RJ. Statistical approaches in landscape genetics: an evaluation of methods for linking landscape and genetic data. Ecography (Cop). Blackwell Publishing Ltd; 2009;32: 818–830.
  112. 112. Spalding MD, Fox HE, Allen GR, Davison N, Ferdaña ZA, Finlayson M, et al. Marine ecoregions of the world: a bioregionalization of coastal and shelf areas. Bioscience. 2007;57: 573–583.
  113. 113. Alvarez R, Farías A, Galvez-Larach M, Hucke-Gaete R, Lo Moro P, Montecinos Y, et al. Sintesis del estudio “Investigación para el desarrollo de Área Marina Costera Protegida Chiloé, Palena y Guaitecas.” In: Hucke-Gaete R, Lo Moro P, Ruiz J, editors. Conservando el mar de Chiloé, Palena y Guaitecas. Primera Ed. Valdivia, Chile; 2010. p. 174.
  114. 114. Iriarte JL, González HE, Liu KK, Rivas C, Valenzuela C. Spatial and temporal variability of chlorophyll and primary productivity in surface waters of southern Chile (41.5–43° S). Estuar Coast Shelf Sci. 2007;74: 471–480.
  115. 115. Iriarte JL, González HE, Nahuelhual L. Patagonian fjord ecosystems in southern Chile as a highly vulnerable region: problems and needs. Ambio. 2010;39: 463–466. pmid:21090000
  116. 116. Shulzitski K, Sponaugle S, Hauff M, Walter KD, Cowen RK. Encounter with mesoscale eddies enhances survival to settlement in larval coral reef fishes. Proc Natl Acad Sci. 2016;113: 6928–6933. pmid:27274058
  117. 117. Balbontín F, Bernal R. Distribución y abundancia del ictioplancton en la zona austral de Chile. Cienc y Tecnol del Mar. 1997;20: 155–163.
  118. 118. Bernal R, Balbontín F. Ictioplancton de los fiordos entre el golfo de Penas y estrecho de Magallanes y factores ambientales asociados. Cienc y Tecnol del Mar 1999;22: 143–154.
  119. 119. Dávila PM, Figueroa D, Müller E. Freshwater input into the coastal ocean and its relation with the salinity distribution off austral Chile (35–55°s). Cont Shelf Res. 2002;22: 521–534.