PREDICTION AND GENETIC PARAMETERS OF FILLET TRAITS IN THREE NILE TILAPIA Oreochromis niloticus STRAINS BASED ON THEIR BODY MEASUREMENTS

Kassaye Balkew Workagegnab*, Gunnar Klemetsdala, Elias Dadeboband Hans Magnus Gjøena
aDepartment of Animal and Aquacultural Sciences, NUNMB, P.O. Box 5003, N-1432 Ås, Norway
bDepartment of Biology, Hawassa University (HwU), P.O. Box 05, Hawassa, Ethiopia
*kassayebalkew@gmail.com
 

The aim of this study was to develop non-invasive methods to predict fillet traits, i.e. fillet weight and fillet yield, in Nile tilapia (Oreochromis niloticus) using their body measurements, by recording 958 male fish from 81 full-sib families. A simple linear regression model utilising body measurements on live fish was first established by forward selection procedures, and then cross-validated, 5-fold, to estimate the prediction power of the model. The best final prediction models for the two fillet traits both had two variables; body weight and body thickness for fillet weight, and body thickness and head thickness for fillet yield. The fillet weight model explained 94% of the observed variance, with a 4.0g root mean standard error prediction (RMSEP) and 0.02g prediction bias, while only 21% of the observed variance for fillet yield was explained by the model, with a 2.3% RMSEP and only 0.002% prediction bias. The estimates of heritability for fillet weight (0.17) and predicted fillet weigh (0.19) were moderate, and had a high genetic correlation (0.94). The estimate of heritability for actual fillet yield was low (0.09) whereas it was higher for predicted fillet yield (0.26), but the two traits had a relatively low genetic correlation (0.50). In conclusion, selection for faster growth of Oreochromis niloticus is likely to result in a significant correlated response in fillet weight, whereas it is likely to only give a minor response for fillet yield.