Exploiting genomic data to estimate effective population size from linkage disequilibrium and to identify genomic regions involved in sex determination in Spanish Atlantic salmon

Advances in Atlantic salmon genomics in recent years have led to the development of high-density single nucleotide polymorphisms (SNP) chips opening thus new opportunities for investigating the genetic dynamics of populations. In particular, they can be used to infer ancestral and current population sizes and for detecting genes affecting life history traits. In this study, the 220K high-density Affymetrix SNP genotyping array (Aquagene-CIGENE) has been used to characterise patterns of linkage disequilibrium, to estimate effective population size from these patterns, and to identify genomic regions involved in sex determination in Spanish Atlantic salmon populations. These populations suffer the most extreme conditions of the distribution range of the species in the world. Samples from six rivers (Miño, Ulla, Eo, Sella, Urumea and Bidasoa) covering all the distribution area of the species in Spain were genotyped. After quality control, 187 fish and more than 150,000 SNPs were available for the analyses. Linkage disequilibrium, measured by the squared correlation coefficient between SNP pairs, was found to be relatively high between close markers (more than 0.5 for markers 0.005 Mb apart), although it declined rapidly with increasing distance (decreasing by 90 per cent at 0.3 Mb). The population structure analysis revealed that two main ancestral clusters compose the current Spanish population. These clusters include an Atlantic group with individuals from rivers Miño and Ulla (Atlantic metapopulation), and a Cantabric group with individuals from rivers Sella, Urumea and Bidasoa (Cantabric metapopulation). River Eo fell between both groups. The Atlantic metapopulation showed higher levels of disequilibrium and a smaller population size than the Cantabric metapopulation. Across rivers, estimates of effective population size ranged from about 500 (Miño) to 1200 (Bidasoa) individuals 50 generations ago, and from 100 to 400 individuals six generations ago. Through a genome-wide association study a total of 317 SNPs were found to be significantly associated with sex determination. This translated into nine putative QTL regions on six chromosomes (Ssa02, Ssa06, Ssa09, Ssa10, Ssa21 and Ssa22). The functional annotation revealed that the most interesting regions associated with sex differentiation and sexual development were in chromosomes Ssa02 and Ssa06 because they may contain the following genes: (i) the sdY gene, which is the master-sex-determining gene already identified in 15 species of salmonids (Ssa02, at 36 Mb, very close to one of our region 2); (ii) the gata4 gene, which encodes a transcription factor involved in gonads development (Ssa06, at 63 Mb) ; (iii) the rspo1 gene, which produces a secreted activator protein in ovary (Ssa06, at 63 Mb); and (iv) the esr1 gene, an estrogen receptor gene which is essential for sexual development and reproductive function in females (Ssa06, at 66 Mb, close to another of our regions). To our knowledge, this is the first study identifying genomic regions associated to sex determination in this species and investigating the magnitude of effective population size from genomic data