Widespread genetic introgression of escaped farmed Atlantic salmon in wild salmon populations

ICES Journal of Marine Science (2016), 73(10), 2488–2498. doi:10.1093/icesjms/fsw121 Editor’s Choice Widespread genetic introgression of escaped farmed Atlantic salmon in wild salmon populations Sten Karlsson*,‡, Ola H. Diserud‡, Peder Fiske, and Kjetil Hindar *Corresponding author: tel: þ47 91124058; fax: þ4773851401; e-mail: ‡ Shared first authorship. Karlsson, S., Diserud, O. H., Fiske, P., and Hindar, K. Widespread genetic introgression of escaped farmed Atlantic salmon in wild salmon populations. – ICES Journal of Marine Science, 73: 2488–2498. Received 7 March 2016; revised 8 June 2016; accepted 15 June 2016; advance access publication 22 July 2016. Farmed Atlantic salmon (Salmo salar) escape from net pens and enter rivers to spawn, potentially resulting in genetic introgression and reduced fitness of wild salmon. Here, we quantify genetic introgression of farmed to wild salmon, using molecular genetic markers, in populations from 147 salmon rivers, representing three-quarters of the total wild salmon spawning population in Norway. For 109 rivers with adult modern samples and sample sizes of 20 or more, the average level of farmed genetic introgression was 6.4% (median ¼ 2.3%), with a range between 0.0% and 42.2%. Fifty-one of these rivers showed significant farmed genetic introgression when compared with historical reference samples. We observed a highly significant correlation between estimated farmed introgression and average proportion of escaped farmed salmon. We quantify levels of introgression as unweighted averages or weighted by population sizes, to compare geographical regions and to compare levels of introgression in rivers and fjords designated as locations deserving a high level of protection. We found a generally lower level of introgression in National Salmon Rivers and National Salmon Fjords subjected to formal protection by parliament. We conclude that farmed to wild genetic introgression is high in a large proportion of Norwegian salmon rivers, with the highest levels found in the most intensive areas of salmon farming. The extensive genetic introgression documented here poses a serious challenge to the management of farmed and wild Atlantic salmon in Norway and, in all likelihood, in other regions where farmed-salmon escape events occur with regularity Keywords: atlantic salmon, aquaculture, farmed salmon, genetic introgression, genetics, SNPs. Introduction Farmed Atlantic salmon differ genetically from wild salmon because of a variety of causes. Breeding programs of farmed Atlantic salmon were established in Norway in the early 1970s based on salmon collected from several populations in Central and Western Norway (Gjedrem et al., 1991; Gjøen and Bentsen, 1997). The breeding program has successfully changed the genetics of farmed Atlantic salmon to improve commercially important traits, such as growth, utilization of feed, and filet quality (Thodesen et al., 1999; Gjedrem and Baranski, 2009; Solberg et al., 2013). These genetic improvements have undoubtedly contributed to the rapid expansion of the Atlantic salmon farming industry in Norway, with a production close to 1.3 million tons in 2015. Farmed Atlantic salmon also differ genetically from wild salmon because of selection to captivity, and loss of genetic variation from a limited number of wild founders and subsequent genetic drift (Hutchings and Fraser, 2008). Because of the reduced fitness (Fleming et al., 2000; McGinnity et al., 2003; Skaala et al., 2012; Reed et al., 2015) and lower genetic variation in farmed salmon (Mjølnerød et al., 1997; Skaala et al., 2004, 2005; Karlsson et al., 2010) compared with their wild conspecifics, there is a concern that genetic introgression of escaped farmed salmon to wild salmon might reduce the viability of wild Atlantic salmon. Reported numbers of escaped farmed salmon in Norway have ranged from 39 000 to 920 000 since 1993, with an average of 380 000 (Norwegian Directorate of Fisheries, http://www.fiskeridir.no/ English). Inventories since 1989 have shown high proportions of C International Council for the Exploration of the Sea 2016. All rights reserved. V For Permissions, please email: Norwegian Institute for Nature Research (NINA), Sluppen, No 7485, P.O. Box 5685, Trondheim, Norway Farmed to wild salmon genetic introgression Material and methods To quantify genetic introgression resulting from spawning of escaped farmed salmon in the wild, we analysed only fish hatched in the wild. We excluded fish classified as escaped farmed salmon, or with uncertain classification, based on their growth patterns in the scales (Lund and Hansen, 1991; Fiske et al., 2005). Samples of juvenile, pre-smolt salmon can safely be regarded as hatched in the wild, because the escape of juvenile farmed salmon from land-based facilities to rivers in this study is unlikely. We extracted total genomic DNA from scales of adult salmon and from fin-clips of juvenile salmon using DNEASY tissue kit (QIAGEN). Initially, we used the Sequenom SNP-genotyping platform for genotyping of 5897 individuals at 99 SNP loci, with PCR amplifications in 4 multiplexes. Primer extension reactions followed recommendations from Sequenom (www.sequenom. com) and fragments were separated and identified using Sequenom Mass ARRAYTM analyzer (Autoflex mass spectrometer). We conducted genotyping in real time depending on the presence or absence of a mass peak in expected mass range for each locus (Tang et al., 1999) using the MassARRAYTM RT 3.4 software. We obtained reliable genotypes from 59 SNPs described as being collectively diagnostic in differentiating between wild and farm salmon (Karlsson et al., 2011; Jensen et al., 2013). For the remaining 15 293 individuals, we used the EP1TM 96.96 Dynamic array IFCs genotyping platform (Fluidigm, San Francisco, CA). Reliable genotypes were obtained for 48 of the same SNPs genotyped by the Sequenom platform (Karlsson et al., 2011). The SNP genotypes from the Sequenom and the Fluidigm SNP genotypes were merged for the 48 common SNP loci (Supplementary Table S1). As a reference for farmed salmon, we used genotypes from 503 individuals from the three leading breeding companies (Marine Harvest, Salmobreed and AquaGen) from the year classes 2004–2009 (MH), 2004–2007 (SB) and 1998–2001, 2008 (AG). Each yearclass represented one of four different breeding kernels from each breeding company. In 2005, AquaGen pooled the four breeding kernels into one big kernel, represented by the 2008 sample. To investigate historical genetic signatures of farmed salmon, we used 129 samples from 1982 to 1988 from the four AquaGen breeding kernels. As references for nonadmixed wild salmon, we used historical samples of 2187 wild individuals from 39 populations, geographically distributed in rivers from southern to northern Norway. In agreement with previous studies (Bourret et al., 2013; Jensen et al., 2014), the Norwegian populations clustered into an Atlantic and a Barent (...truncated)