Studies of population and ecological genetics are rapidly moving into ecological genomics with the advent of genotyping using affordable high-density sequencing (genotyping-by-sequencing or GBS). Our historical emphasis, using single nucleotide polymorphisms (SNPs) to study life history, migration, and improved management of Pacific salmon, naturally morphed into ecological genomics with access to the new datasets acquired through GBS (Seeb et al. 2011a). Our lab still pursues population studies that couple SNP variation with life history variation and migration; these studies use high-throughput genotyping of batteries of 96 SNPs in thousands of individuals. But the SNP discovery pipelines using GBS have matured to the point that we can now do genomics studies using thousands of SNPs that span the genomes of 100s of individuals.
Investigation of life history variation
We work closely with the University of Washington’s Alaska Salmon Program which conducts long-term research on populations of salmon in a series of field camps in Bristol Bay and Southwest Alaska. A major focus is on sockeye salmon, a charismatic species of iconic value for nations that surround the North Pacific Ocean and Bering Sea. A large portion of our funding is directed towards elucidating population structure and studying factors that shape ecological and life history variation.
We collect phenotypic measurements on sockeye salmon to relate life history variation to genomic results.
Recent studies in Bristol Bay demonstrated relationships among genetic structure, smolt migration timing, and spawning habitat heterogeneity (McGlauflin et al. 2011; Ruff et al. 2011). Gomez-Uchida et al. (2011) identified effects of temporal isolation that were a greater driver of population structure than were those of geographic distance. Creelman et al. (2011) documented similar findings in populations that inhabit lakes in the Chignik River drainage in Southwest Alaska. All of these studies were strongly influenced by allelic frequency differences observed for outlier loci that encode proteins such as major histompatiblity complex or transferrin—loci that may be candidates for natural selection.
Uncertainty about the inter-decadal stability of allelic frequencies at outlier loci concerns some who criticize the use of long-term DNA datasets to make contemporary decisions for conservation and management. Most of the large datasets for salmonids are assembled from samples collected over many decades (e.g., Habicht et al. 2010; Seeb et al. 2011d; Templin et al. 2011); is it possible that allelic frequencies for outlier loci could change during this time span? We investigated the stability of allele frequencies over 25 – 42 years (approx. 5 – 8 generations) in sockeye salmon from three Alaska systems (Gomez-Uchida et al. 2012). We found that temporal changes were dramatically smaller (between 40- and 250-fold) than spatial differences in allele frequencies based on 88 SNPs including outlier loci described above. The magnitude of temporal change was consistent with a model of genetic drift, ameliorating concerns for the use of long-term datasets.
Migration of Pacific salmon on the high seas
Scientists and crew from R/V Oshoro maru set a net for high-seas collections.
Pacific salmon (Oncorhynchus sp.) undertake long migrations across the North Pacific Ocean and through the Bering Sea, with populations originating from North America and Asia often sharing migratory routes and feeding grounds during their high-seas residencies. Pacific salmon are anadromous, generally returning to their natal freshwater origins to spawn, forming hierarchical meta-population structures. We use SNP datasets to investigate this structure and to track the migration of immature fish on the high-seas and near-shore environment. These applications are particularly timely with increasing interest in the effects of climate change and the role of increasing hatchery production on migration of salmon. Our recent studies, in collaboration with scientists from Japan, Russia, and Korea, focused on developing range-wide genetic databases and tracking migratory pathways for sockeye, chum, and Chinook salmon (Habicht et al. 2010; Seeb et al. 2011c; Seeb et al. 2011d; Templin et al. 2011).
New resources for ecological genomics: hi-density genetic maps for duplicated salmonids
Our early efforts to develop ‘genomic’ resources were largely limited to the sequencing of known expressed sequence tags (ESTs) to discover SNPs (for example see such papers as Smith et al. 2005). As outlier loci became more important in our studies (e.g., Ackerman et al. 2011), we switched to transcriptome sequencing (Seeb et al. 2011b) to discover thousands of novel ESTs. Our goal was to find SNPs in candidate genes that could provide insight into adaptive variation and lead to higher resolution for population studies.
Collection of salmon gametes for experimental crosses used to construct meiotic maps.
SNP discovery in transcriptome was generally confounded by the variability of gene expression and by the presence of intron/exon boundaries (Seeb et al. 2011b; Roberts et al. 2012); SNP discovery in salmonids was further confounded because of the difficulty of sorting polymorphisms between homologs and ohnologs (chromosomes derived from whole genome duplications) in the recently duplicated genome (Everett et al. 2011). At this juncture we realized three things: 1) yes, describing outlier loci would soon start to provide keener insight into adaptive variation, 2) applying the recently developed GBS techniques in genomic DNA provided vast advantages over transcriptome sequencing for simple SNP discovery (e.g., restriction site associated DNA (RAD) sequencing), and 3) dense genetic maps would benefit SNP discovery and validation in the duplicated salmonid genome. It became clear that we were entering a new era where the combination of GBS and dense maps of loci responsible for adaptive variation would naturally transition our laboratory into the study of ecological genomics.
Dense meiotic maps
Linkage group 2 in sockeye salmon (left female, right male). Map lengths measured in salmonid males are typically shorter than those observed in females because of areas of reduced recombination.
Meiotic maps are a critical component of the process of unraveling the complexities of the genomics of duplicated taxa. Pacific salmon are derived from two whole genome duplications that occurred, first in teleosts over 300 MYA, and then in salmonids about 25 MYA. Our first effort, from diploid matings of sockeye salmon, mapped 1772 RAD-tag and EST-based SNPs spread throughout the genome of sockeye salmon. We are now constructing maps resulting from haploid families from four species: chum, pink, and Chinook salmon and steelhead trout. These maps will be used in concert with our life history and ecological studies to pinpoint genes and linkage groups involved in ecologically relevant traits and to help understand the roles those genes play in an organism's function and evolution.