Early life exploration behavior and life-history loci are co-localized in an adaptive genomic hotspot in Atlantic salmon

The traits that are important for adaptation may exhibit genetic correlation due to pleiotropy or as a result of linkage. Understanding the genetic architecture of such correlations is important for predicting the selection response of populations. Exploration in fishes is a behavioral trait by which individuals may find habitats with better foraging and growth opportunities that subsequently improve their fitness. For example, in Atlantic salmon ( Salmo salar) , all individuals originate in spawning rivers where females lay eggs, but some juveniles migrate to small tributaries which are not spawning areas but provide favorable habitat patches for young salmon. The increased growth in these nursery streams may facilitate earlier sexual maturation, implying a potential correlation between exploration and maturation traits.

In this study, by sampling juveniles from two wild populations in the large Teno River catchment in northernmost Fennoscandia, we tested the genetic association between exploration behavior with nursery streams across four SNPs that span a 70-kb long genomic region with a major effect on age at maturity variation. Three of these SNPs are missense mutations in the vgll3 and akap11 genes, and one SNP tags a putative regulatory region with the strongest association with the age-at-maturity trait.

We show that the exploration behavior was linked to the genomic region in one of the two studied populations. However, the genetic association was substantial in the missense SNP located in the akap11 gene, which is farthest away from the vgll3 SNPs and previously ruled out as being linked to the age at maturity. We also detected a marginal interaction effect between SNPs in the vgll3 gene and akap11 , indicating a potentially complex genetic architecture underlying the trait variation.

Our results suggest that exploration and age at maturity are co-inherited within the same haplotype block, but we find no evidence for direct causality. These two traits may form a co-adapted trait complex that may be instrumental in local adaptation processes.