Elsevier

Aquaculture

Volume 253, Issues 1–4, 31 March 2006, Pages 229-237
Aquaculture

Genetic variability in wild and farmed Atlantic salmon (Salmo salar) strains estimated by SNP and microsatellites

https://doi.org/10.1016/j.aquaculture.2005.09.022Get rights and content

Abstract

Wild and farmed stocks of Atlantic salmon were genotyped, using 16 microsatellites and 26 SNPs. Some population parameters were estimated, as well as the power of these markers to assign single individuals to their sampling population. The efficiencies of the marker sets for parentage testing in farmed populations were also estimated.

A total of five wild river strains and two farmed stocks were genotyped. The average heterozygosity was 0.41 and 0.51 for the SNPs and microsatellites, respectively. The microsatellite set displayed an average allele number of 3.8 within populations and 7.9 per locus.

The calculated exclusion probabilities of the parentage analyses based on each of the two marker sets were better than 99.9%, which demonstrates their robustness in parentage testing.

Genetic assignment analysis allocated 95.8% of the individuals to the sampling population. Neighbour-joining phylograms based on individual genotypes from both marker sets tended to cluster individuals according to their sampling population. Some potential strain-specific alleles were discovered for the microsatellites. Such alleles may have a potential additional value of population identification.

The SNP collection showed nearly as good results as the microsatellites in this study and improved high throughput typing methods for SNP detection suggest that they can be a good alternative to microsatellites for future typing.

Introduction

Microsatellites, and recently single nucleotide polymorphisms (SNPs), have been widely used as DNA markers in genome analyses, paternity testing, genetic fingerprinting and calculation of genetic distances in salmonids within the last ten years (Beuzen et al., 2000, Fontaine et al., 1997, Kikuchi and Isagi, 2002, Koskinen et al., 2002a). A few papers focus on the comparison of marker systems in other animal species (Seddon et al., 2005, Vignal et al., 2002), but such comparisons have not so far been reported in salmonids.

The major advantages of employing microsatellites in such studies are their high degrees of polymorphism and their abundance throughout the genomes, spaced at an average physical distance of 30–100 kb in the eukaryotic genome (Hearne et al., 1992). SNPs are the result of mutations leading to nucleotide substitutions or the deletion/addition of individual nucleotides. Since SNP loci usually contain two alleles, genotyping them requires plus/minus assays only, as opposed to microsatellites, where length differentiation of a series of alleles is needed. This makes SNPs suitable for automated analysis using DNA chip technologies, mass spectrometry etc., and are more computer friendly due to their biallelic nature (Landegren et al., 1998, Nielsen, 2000, Pastinen et al., 2000). SNPs occur more frequently in the genome than microsatellites. In the human genome SNPs appear at approximately every 1000–2000 bp (Altshuler et al., 2000, Kwok et al., 1996, Sachidanandam et al., 2001).

Mutation rates in SNPs are significantly lower than in microsatellites (Landegren et al., 1998, Nielsen, 2000). This stability of SNPs is an advantage in evolutionary studies, population biology and pedigree studies. Microsatellites often appear in non-coding regions of the genome, while SNPs can also be detected in the coding sequences. Thus, SNPs may affect protein function and expression levels directly (Landegren et al., 1998) and hence also be subjected to evolutionary selective forces (Stoneking, 2001). Due to their biallelic nature, SNPs are less informative than microsatellites, but their information content can be increased by including more SNP markers (Glaubitz et al., 2003). The information value of a marker set is also influenced by the allele frequencies. For exclusion probability and individual traceability an equal frequency of available alleles is favourable (Glaubitz et al., 2003, Vignal et al., 2002). In general, the value of a certain marker set varies depending on the purpose and testing material, and the set should be tested in advance to assure sufficient power (Vignal et al., 2002).

Molecular genetic markers such as microsatellites and SNPs have a wide range of potential applications for long-term management of farmed and wild salmon populations. The markers are powerful tools for distinguishing between populations (Hansen et al., 2001) or for estimating the impact of farmed salmon on wild stocks (Koskinen et al., 2002b, Seddon et al., 2005, Skaala et al., 2004).

In the present study we compare the power and efficiency of the two genetic marker systems for use in population studies, parentage testing and the genetic assignment of individual salmon.

Section snippets

Biological material

Individuals from seven different populations were collected (Table 1 and Fig. 1). Fin samples from a total of 167 Atlantic salmon were stored in 95% ethanol at − 20 °C and DNA was isolated by standard methods.

The AquaGen population is a commercial breeding stock based on a long-term breeding programme systematically selected for economically important traits. The brood stock was sampled in more than 40 different Norwegian rivers and localities in the early 1970s, and pooled in four

Genotyping results and frequencies of SNPs and microsatellites

Sixty potential SNPs were initially identified in the salmon genome. On the basis of allelic frequency and PCR functionality, 26 of these were selected for analysis in the population. We suggest that the frequency of suitable SNPs in the salmon genome is approximately one in every 3000 bp, estimated by the number of SNPs identified and the number of base pairs sequenced.

Salmon from five rivers (Otra, Figgio, Gaula, Lona, and Spey) and two farmed stocks (AquaGen and Rauma) were typed using the

Discussion

In general both marker sets show an acceptable level of variation in the populations, which makes the collections suitable for population studies as well as for pedigree analysis in farmed stocks. The average level of SNP heterozygosity observed (Ho = 0.41) is satisfactory for these purposes. The level of polymorphism in the microsatellites (Ho = 0.51) and the average allele numbers across and within populations are lower than reported in other studies on Atlantic salmon (Fontaine et al., 1997,

Acknowledgements

We gratefully thank Monica Skogen for her laboratory work on the microsatellites. The project was founded by the Research Council of Norway, contract no. 138792/130.

References (31)

  • J. Felsenstein

    Phylogeny inference package

  • P.M. Fontaine et al.

    A genetic test of metapopulation structure in Atlantic salmon (Salmo salar) using microsatellites

    Can. J. Fish. Aquat. Sci.

    (1997)
  • J.C. Glaubitz et al.

    Prospects for inferring pairwise relationships with single nucleotide polymorphisms

    Mol. Ecol.

    (2003)
  • M.M. Hansen et al.

    Assigning individual fish to populations using microsatellite DNA markers

    Fish Fish.

    (2001)
  • A. Jamieson et al.

    Comparisons of three probability formulae for parentage exclusion

    Anim. Genet.

    (1997)
  • Cited by (0)

    View full text