Evaluation of statistical models for genetic analysis of challenge test data on furunculosis resistance in Atlantic salmon (Salmo salar):: Prediction of field survival

Challenge test data of Atlantic salmon for survival to Aeromonas salmonicida infection were used in genetic analyses, by defining survival as either a binary trait (test period survival), repeated measures of a binary trait (test-day survival), or time until death. Test period survival was analyzed by a linear and a threshold model, while a linear repeatability model was used for analysis of test-day survival, and a proportional hazards Weibull model was applied to time until death. The threshold and the proportional hazards models resulted in the highest heritabilities (0.59 and 0.63, respectively), while the lowest heritability was found for the linear repeatability test-day model (0.02). The different models were ranked according to their ability to predict full-sib family survival to disease outbreak in the field. All models proved to be good predictors, with correlations between predicted full-sib family effects from challenge test and full-sib family field survival ranging from 0.71 to 0.75. The linear repeatability model had the best predictive ability, followed by the cross-sectional threshold model, the Weibull frailty model, and the cross-sectional linear model. The Weibull frailty model was unable to include common environmental family effects in addition to additive genetic effects of sire and dam, possibly reducing the predictive ability of this model. However, a high Spearman correlation coefficient (0.98) was found between the predicted breeding values of the Weibull and the linear repeatability models, both defining survival as a longitudinal trait. If the endpoint of recording was set at 50% mortality, Speannan correlations among predicted breeding values from the different models were all very high (> 0.98). (c) 2006 Elsevier B.V. All rights reserved.