4.2 | Genome environment association
Fish employing broadcast spawning strategies characterized by larval and juvenile pelagic drift in ocean currents are subject to large interannual variability in oceanic conditions (Stockhausen et al., 2018). Stockhausen et al. (2018) refers to this as “running the gauntlet”, as it is during this critical life stage that these fish are most vulnerable, experiencing the highest rates of mortality. This vulnerability is not only due to the vagaries of physical transport, but also due to their physiological condition where they must meet energetic demands of acquiring sufficient lipid reserves in order to move to inshore nursery areas.
During years favorable ocean conditions with ample food availability, such as 2014 for POP, mortality may be low and selection weak, allowing most phenotypes to survive through the pelagic phase and into nearshore settlement. However, during years of unfavorable ocean conditions, such as the unusual warming, low primary productivity, and low food availability in 2015 for POP, mortality may be high. If this increase in mortality is especially high for certain phenotypes, the selection may be strong, with only the most favorable phenotypes surviving to settlement.
Our results show consistent selective forces along the sampling date/latitude gradient in both 2014 and 2015 for POP with 10 of the 381 putative selective loci being in common in both years (Table 4). The LFMM analysis was done independently for each of the years and finding the same putative selected loci in both years is surprising. And although the LFMM method purportedly accounts for demographic factors such as population mixtures, the date/latitude gradient association could be due adult spawning populations being differentiated at these loci. Based on timing and location of spawning, their progeny may follow the spatio-temporal pattern identified by GEA. This is further supported by the distribution of the sNMF identified genetic clusters in relation to their distribution as seen in Figure 3. Alternatively, this may indicate that the spawning adult populations contain a high proportion of alleles at those loci that in 2014 and 2015 years were deleterious to the YOY progeny encountering the environmental conditions during their pelagic developmental stage. Since POP are very long-lived and may even spawn into their 100th year (Conrath and Knoth, 2013; Hulson et al., 2017; Heppell et al., 2010), some of the alleles in the parental population are expected to have been selectively advantageous during their respective first year at sea; therefore, the alleles that were advantageous when the parents were YOY may be deleterious in some oceanic conditions encountered by their progeny decades later. It is then expected that patterns of selection as displayed by the subsets of selected alleles would be cohort-specific.
Interannual differences in the strength of selection pressure was evident when comparing the 2014 and 2015 YOY POP fish. Due to the larger sample size in 2014 (321) than in 2015 (77), we would expect more putative selected loci in 2014 just due to the increase in statistical power, but that was not the case. In 2014, the oceanic conditions were typical (Cavole et al., 2016), with large YOY abundances in the ocean, and no putative selected loci were identified aside from those associated with collection date/latitude. However, in 2015, the oceanic conditions were abnormal with warmer sea surface temperatures (Gentemann et al., 2017) and were marked by large seabird die-off (Jones et al. 2018). This likely resulted in stronger selective pressure on YOY in 2015 and this is supported by the greater number of putative selected alleles. Therefore by the time the 2014 and 2015 cohorts settled out in the nearshore, we expect that most individuals have gene variants that were most favorable and selected for by the conditions encountered in that year.
The difference in the change in condition index indicates different growth conditions between the two years. In 2015, the smaller fish had less mass than in 2014, but the larger fish had equivalent mass in both years (Figure 4). This indicates that in 2015, a much warmer year than in 2014, the smaller fish were unable to gain weight as compared to the same sized fish in 2014. If the temperatures were still within optima for POP YOY growth, then smaller size suggests smaller-sized prey items were either unavailable or of insufficient nutritional value to support the higher growth rates predicted by the higher temperature in 2015. However the larger fish in both years were equally successful at gaining mass. This suggests that the environment in 2015 imposed a larger variance in fitness and therefore much stronger selection pressure, and this is consistent with the greater number of putative selected loci in 2015 than in 2014. This is further supported by the recruitment estimates in the 2017 stock assessment with 2014 cohort being 87.5 million and 38.2 million in 2015 (Hulson et al., 2017).