Kokanee–sockeye salmon hybridization leads to intermediate morphology and resident life history: implications for fisheries management

Abstract

Stocking programs designed to return extirpated species to their historical range have become increasingly prevalent, punctuating the need to better understand the risks posed to recipient ecosystems. Here, we investigated the genetic and biological consequences of an anadromous sockeye salmon (Oncorhynchus nerka) stocking program in Skaha Lake, British Columbia, where substantial levels of hybridization and introgression with the native freshwater resident ecotype (kokanee) have been detected. We genetically assigned 543 individuals (adult spawners, age-0 juveniles) to estimate stock proportions (pure-stock sockeye, pure-stock kokanee, or hybrid) between 2010 and 2017, with a subset undergoing otolith microchemistry analysis to determine migratory life history and maternal ancestry. The proportion of hybrid spawners varied from 5% to 20% across sampling years, while that of hybrid age-0 juveniles remained relatively constant (∼11%). Hybrid spawners exhibited intermediate size relative to pure stocks, with the vast majority being nonanadromous (92%) and of resident maternal ancestry (76%). Our results provide empirical support for previously hypothesized mechanisms of hybridization between O. nerka life-history forms and underscore the importance of continued monitoring of stocking programs to quantify long-term fitness impacts of introgression and refine management strategies.

Genotyping-in-Thousands by sequencing panel development and application for high-resolution monitoring of introgressive hybridization within sockeye salmon

“…This program provides an excellent opportunity to investigate the extent and outcomes of intra-specific hybridization between O. nerka migratory and resident forms outside of a laboratory setting. Previous research examining hybridization between sockeye salmon and kokanee found O. nerka hybrids occupy intermediate morphologies and exhibit a largely resident life history, although this latter finding requires further investigation 12 , 29 . These studies employed genetic marker sets that were effective at differentiating pure-stock from hybrid, however, they had limited power for the identification of advanced hybrid classes that has become increasingly important for genetic monitoring as the reintroduction program matures.…”

“…We genotyped tissue samples from O. nerka spawners in the Okanagan River upstream of Skaha Lake that were previously sampled in 2016 (n = 45) and 2017 (n = 59) 29 , and newly sampled in 2019 (n = 218; this study) as deadpitch (post-mortem carcass collection after spawning) by personnel from the BC Ministry of Forests, Lands, Natural Resource Operations, and Rural Development. The section of river that was sampled represents the only spawning habitat that exists for O. nerka populations from Skaha Lake and all forms spawn within this section of river.…”

“…We constructed a GT-seq test library with the previously extracted DNA samples collected in 2016 (n = 45) and 2017 (n = 59) for which otolith microchemistry analysis was previously conducted 29 . Extracted DNA was quantified with a Qubit 3.0 Fluorometer and the dsDNA High Sensitivity Assay Kit (Invitrogen).…”

“…We then evaluated the accuracy, efficiency, and power of the optimized GT-seq panel using simulation analyses. We further assessed panel performance by re-genotyping individuals sampled in Skaha Lake in 2016 and 2017 that were previously genetically assigned as pure-stock or F 1 hybrid at 32 SNPs and subjected to otolith microchemistry analysis to reconstruct migratory history 29 . Finally, we genetically assigned O. nerka spawners sampled in 2019 to pure-stock or hybrid class (F 1 , F 2 , B 2 backcross-kokanee, B 2 backcross-sockeye) to estimate stock proportions and examine trends in this system over time.…”

Genotyping-in-Thousands by sequencing panel development and application to inform kokanee salmon (Oncorhynchus nerka) fisheries management at multiple scales

“…The optimal value of K was then combined across ten independent runs with CLUMPP [ 45 ]. We also assessed individual assignment accuracy with the Rannala and Mountain (1997) estimator as implemented in ONCOR with 100 simulated individuals per population based on empirical genotype frequencies with recom-sim [ 46 – 48 ]. After assessment, we sent the full RAD tag sequences associated with the pool of candidate SNPs to GTseek LLC ( https://gtseek.com/ ) for custom locus-specific primer design.…”