Effects of triploidy on genetic gains in a rainbow trout (Oncorhynchus mykiss) population selectively bred for diploid growth performance

A monosex female, pedigreed rainbow trout population was selected as diploids to improve growth performance to the standard ~500-gram US market weight and beyond (Select line), and a contemporary, randomly-mated control line was maintained to empirically estimate selection response (Synthetic Control line). After two generations of selection, 10 Select and 7 Synthetic Control families were evaluated for growth performance as diploids and pressure-induced triploids with the aim of evaluating effects of triploidy on genetic gains made by selecting on diploid performance. Approximately 24 fish per ploidy × family subclass (810 fish total) were tagged at 5 months of age, commingled into one of three replicated tanks, and measured monthly for body weight, fork length, and condition factor until 21 months of age. Data were analyzed separately by month using a mixed-effects linear model to evaluate fixed effects of genetic line, ploidy, tank, genetic line × ploidy, and genetic line × ploidy × tank, and random effects of family nested within genetic line and family nested within ploidy. By 7 months of age and throughout the remainder of the study, Select-line fish and diploids were heavier compared to Synthetic Control fish and triploids (P ≤ .03). Select-line fish and diploids had numerically larger condition factors throughout most of the study, but differences were generally not significant (P < .05) until the latter 4–6 months of the study. Genetic line did not interact with ploidy to affect body weight (P ≥ .16) or condition factor (P ≥ .36) at any time between 5 and 21 months of age. Between 8 and 21 months of age, body weights of the Select-line fish averaged 22.7% larger than Synthetic Control fish as diploids and 24.5% larger as triploids. Family interacted with ploidy to affect body weight (P ≤ .038) at all times between 5 and 13 months of age, but the amount of phenotypic variation attributable to this interaction tended to decrease over time. By 13 months of age, when body weight approached 1 kg, the interaction accounted for only 7.8% of the phenotypic variation. This study demonstrates that selection on diploid growth performance is effective for improving triploid growth performance, and provides evidence that triploid performance may be further improved by including triploid phenotypes in the breeding objective.