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

Using cross-sectional models, moderate to high heritabilities were estimated for ISA resistance both on the observable scale, using a linear model (0.21), and on the underlying scale, using a threshold model (0.32). Estimated test-day heritability was low (0.01) on the observable scale (linear model), but moderate (0.16) on the underlying scale (logit model). Relatively small differences were observed with respect to predictive ability for the different models. Still, the results suggested that test-day models for survival may improve accuracy of selection by 10–12% over what can be obtained by cross-sectional models.