Evaluating adaptive divergence between migratory and nonmigratory ecotypes of a salmonid fish, Oncorhynchus mykiss

Abstract

Next-generation sequencing and the application of population genomic and association approaches have made it possible to detect selection and unravel the genetic basis to variable phenotypic traits. The use of these two approaches in parallel is especially attractive in nonmodel organisms that lack a sequenced and annotated genome, but only works well when population structure is not confounded with the phenotype of interest. Herein, we use population genomics in a nonmodel fish species, rainbow trout (Oncorhynchus mykiss), to better understand adaptive divergence between migratory and nonmigratory ecotypes and to further our understanding about the genetic basis of migration. Restriction site-associated DNA (RAD) tag sequencing was used to identify single-nucleotide polymorphisms (SNPs) in migrant and resident O. mykiss from two systems, one in Alaska and the other in Oregon. A total of 7920 and 6755 SNPs met filtering criteria in the Alaska and Oregon data sets, respectively. Population genetic tests determined that 1423 SNPs were candidates for selection when loci were compared between resident and migrant samples. Previous linkage mapping studies that used RAD DNA tag SNPs were available to determine the position of 1990 markers. Several significant SNPs are located in genome regions that contain quantitative trait loci for migratory-related traits, reinforcing the importance of these regions in the genetic basis of migration/residency. Annotation of genome regions linked to significant SNPs revealed genes involved in processes known to be important in migration (such as osmoregulatory function). This study adds to our growing knowledge on adaptive divergence between migratory and nonmigratory ecotypes of this species; across studies, this complex trait appears to be controlled by many loci of small effect, with some in common, but many loci not shared between populations studied.

Keywords: SNP; genomics; life history variation; salmonids; smoltification.

Figure 1

F_ST for markers that were mapped to two linkage maps of the O. mykiss genome (Miller et al. 2012; Hecht et al. 2012). Markers in red represent significant outliers as determined by LOSITAN (for details, see the section Materials and Methods), black line represents the kernel smoothed distribution of all markers (both outliers and “neutral” loci). Chromosomes are marked by alternate shading with a some of the chromosomes labeled. (A) F_ST between Sashin Creek residents and Sashin Creek migrants and (B) F_ST between Little Sheep Creek residents and Little Sheep Creek migrants.

Figure 2

Tajima’s D results for markers that mapped to two linkage maps of the O. mykiss genome (Miller et al. 2012; Hecht et al. 2012). Markers in blue produced a significant positive Tajima’s D and markers in red a significantly negative Tajima’s D. The black line represents the kernel-smoothed average Tajima’s D for all mapped markers. Chromosomes are alternately shaded with a subset labeled. (A) Tajima’s D calculated within the Sashin migrants, (B) within the Little Sheep Creek samples (combined migrants and residents).

Figure 3

The kernel-smoothed average observed heterozygosity in markers, which mapped to two linkage maps of the O. mykiss genome (Miller et al. 2012; Hecht et al. 2012). The kernel-smoothed average for all markers that mapped is presented for the (A) Sashin migrants, (B) Sashin residents, and (C) Little Sheep Creek samples. Chromosomes are alternately shaded.

Figure 4

Nucleotide diversity as calculated by Watterson’s theta in markers that mapped to two linkage maps of the O. mykiss genome (Miller et al. 2012; Hecht et al. 2012). The kernel-smoothed average for all markers that mapped is presented for (A) Sashin Creek migrants and (B) Little Sheep Creek samples. Chromosomes are alternately shaded with some chromosomes labeled.