Genomic vulnerability of a freshwater salmonid under climate change

Abstract

Understanding the adaptive potential of populations and species is pivotal for minimizing the loss of biodiversity in this era of rapid climate change. Adaptive potential has been estimated in various ways, including based on levels of standing genetic variation, presence of potentially beneficial alleles, and/or the severity of environmental change. Kokanee salmon, the non-migratory ecotype of sockeye salmon (Oncorhynchus nerka), is culturally and economically important and has already been impacted by the effects of climate change. To assess its climate vulnerability moving forward, we integrated analyses of standing genetic variation, genotype-environment associations, and climate modeling based on sequence and structural genomic variation from 224 whole genomes sampled from 22 lakes in British Columbia and Yukon (Canada). We found that variables for extreme temperatures, particularly warmer temperatures, had the most pervasive signature of selection in the genome and were the strongest predictors of levels of standing variation and of putatively adaptive genomic variation, both sequence and structural. Genomic offset estimates, a measure of climate vulnerability, were significantly correlated with higher increases in extreme warm temperatures, further highlighting the risk of summer heat waves that are predicted to increase in frequency in the future. Levels of standing genetic variation, an important metric for population viability and resilience, were not correlated with genomic offset. Nonetheless, our combined approach highlights the importance of integrating different sources of information and genomic data to formulate more comprehensive and accurate predictions on the vulnerability of populations and species to future climate change.

Keywords: Pacific salmon; climate vulnerability; local adaptation; standing variation; structural variants.

FIGURE 1

Map of sampled kokanee populations and maximum likelihood phylogenetic tree. Each point in the map represents a lake sampled for this study, and the background represents recent (1970–2000) climate data for bio5, the maximum air temperature of the warmest month. On the left side is the ML tree, constructed based on population allele frequencies, with individuals divided by ecotype in populations where stream‐ and shore‐spawning kokanee co‐occur.

FIGURE 2

Principal component analysis (PCA) plots showing population differentiation based on sequence (a, b) and structural (c, d) variation along PCA axes 1 and 2 (a, c) and axes 1 and 3 (b, d).

FIGURE 3

Histograms of the R ² weighted importance of environmental variables in explaining adaptive variation. Strong sequence outliers only on the left (a) and all structural outliers on the right (b).

FIGURE 4

Adaptive index plots calculated from sequence variation (top) and structural variation (bottom) based on RDA axis 1 (left), most strongly associated with temperature variables, and axis 2 (right), most strongly associated with precipitation variables. Circles represent sampled locations.

FIGURE 5

Top, genomic offset estimates based on sequence (a, c) and structural variation (b, d) according to the best‐case (RCP2.6; a, b) and worst‐case climate change scenarios (RCP8.5; c, d). The map background shows changes in the maximum temperature of the warmest month (bio5 ΔT) predicted for the period 2041–2060. Below, plots showing the correlation, or lack thereof, of genomic offset estimates with latitude (e), bio5 ΔT (f), and genomic heterozygosity (g).