Single‐parentage assignments reveal negative‐assortative mating in an endangered salmonid

Abstract

Understanding reproductive patterns in endangered species is critical for supporting their recovery efforts. In this study we use a combination of paired‐parent and single‐parent assignments to examine the reproductive patterns in an endangered population of sockeye salmon ( Oncorhynchus nerka ) that uses Redfish Lake in central Idaho as a spawning and nursery lake. Recovery efforts include the release of maturing adults into the lake for volitional spawning. The lake is also inhabited by a population of resident O . nerka that is genetically indistinguishable, but phenotypically smaller, to the maturing adults released into the lake. The resident population is difficult to sample and the reproductive patterns between the two groups are unknown. We used results of paired‐ and single‐parentage assignments to specifically examine the reproductive patterns of male fish released into the lake under an equal sex ratio and a male‐biased sex ratio. Assignment results of offspring leaving the lake indicated a reproductive shift by males under the two scenarios. Males displayed an assortative mating pattern under an equal sex ratio and spawned almost exclusively with the released females. Under a male‐biased sex ratio most males shifted to a negative‐assortative mating pattern and spawned with smaller females from the resident population. These males were younger and smaller than males that spawned with released females suggesting they were unable to compete with larger males for spawning opportunities with the larger, released females. The results provided insights into the reproductive behavior of this endangered population and has implications for recovery efforts.

Inferring precocial Chinook Salmon production through single‐parentage assignments

“…Single-parentage analyses have been used to provide insights into salmonid systems (Shedd et al 2022;Steele et al 2022). However, conducting single-parentage analyses can be prone to misassignment, and researchers often find ways to corroborate single-parentage assignments or resort to post hoc approaches to ensure confidence in a single-parentage assignment.…”

“…However, conducting single-parentage analyses can be prone to misassignment, and researchers often find ways to corroborate single-parentage assignments or resort to post hoc approaches to ensure confidence in a single-parentage assignment. Various approaches exist for conducting single-parentage analyses, but we find that an elegant solution for implementing single-parentage assignments is to evaluate a data set's ability to accurately make assignments and then select an appropriate threshold that allows for acceptable levels of false-positive and false-negative error rates (Steele et al 2022). This allows researchers to estimate a project-specific error rate that can be controlled through selection of the assignment threshold.…”

“…Single-parentage analysis followed the protocols described by Steele et al (2022) and was conducted using the R package gRandma (https:// github. com/ delom ast/ gRandma).…”

“…To estimate an analysis-wide error, we first used COLONY (Jones and Wang 2010) to identify groups of full siblings within the parental data set by using the program's full-likelihood analysis method under a polygamous mating system, but otherwise, the default settings were used: no inbreeding (as recommended by the COLONY user guide), a medium-length run, and without sibship size priors. Using the methods presented by Steele et al (2022), we then used the results to estimate the probability that a single-parentage comparison in the analysis would contain the full sibling of the true parent (P fs ), and we incorporated this probability into an estimate of the experiment-wide false-positive error rate (F p ). In this study, due to high weir efficiency, we essentially sampled the entire population of spawning adults to estimate P fs .…”

“…Most parentage analyses assume that no extended family structure exists within a data set, and the inclusion of candidate parents that are related to one another violates this assumption. Assignment accuracy can be improved as more markers are used (Harrison et al 2013b), but explicitly reporting on how the selection of an assignment threshold can vary within the presence of related parents is still warranted (Steele et al 2022).…”