Breeding for improved production of Nile tilapia (Oreochromis niloticus L.)

Summary

The aim of this thesis was to generate knowledge that supports the design of breeding programs for Nile tilapia targeting genetic improvement of body weight and fillet yield to serve the European market. To this end, both the genetic variation and the performance levels of different strains of tilapia were investigated within intensive recirculation systems. Molecular genetic variation was investigated by use of microsatellite markers on four domesticated strains of tilapia. Moderate to great genetic differentiation was found and all strains had added value for the total genetic diversity.To enable phenotypic selection for fillet weight- and yield based on body measurements, phenotypic relationships between body measurements and fillet weight and fillet yield were investigated. Strong, almost linear, relationships between body measurements and fillet weight were found, but relationships with fillet yield were weak. Prediction models for fillet weight were accurate (R ² =0.95) but not for fillet yield (R ² =0.38) meaning that phenotypic selection for fillet weight is possible but hard for fillet yield. Genetic parameters for body weight, fillet weight,filletyield and body measurements were estimated to evaluate the potential of mass selection for fillet traits based on body measurements. The accuracy of a selection index including only body weight indicated that in this way almost the same amount of the selection response can be achieved compared to what hypothetical direct selection for fillet weight would. The use of only width in the selection index would result in 8.5% lower selection response than the use of body weight. It was concluded that body weight is the best predictor for fillet weight compared to body measurements.Genetic parameters for body weight at all ages ranging from 100 days to 326 days were estimated using a random regression model with covariance functions. The heritability of body weight was fairly constant around 0.2, which offers good prospects for selection on body weight. Genetic correlations were estimated between all ages and showed that early selection results in higher selection response than direct selection, when the target trait is body weight at the age of 326 days. Furthermore, evidence was found for genetic differences in growth patterns of fish of different strains. Phenotypic differences in body weight between male- and female fish were significant already at early ages (100 days). Results from a bivariate genetic analysis suggested that body weight in male and female fish is most likely controlled by the same genes since genetic correlations were high (>0.85) at all five measurement dates. Prospects to decrease the difference between mature male- and female body weight by selection, to reduce the total variation, are therefore unfavorable in Nile tilapia.New methodology to estimate variance components for genetic effects of competition was applied to Nile tilapia. No genetic variation for competitive behavior was found although environmental variation for competitive behavior could clearly be detected. This methodology is potentially very interesting for breeding in aquaculture, because no observations on behavior are required.Finally, the advantages and disadvantages of three selection strategies are discussed in terms of necessary facilities, workload and selection responses. It was argued that when traits measurable on live animals are targeted for genetic improvement, mass selection offers a cheap and simple strategy that can result in relatively high selection responses. However, when more complicated traits are desired, both family- and BLUP selection offer better possibilities depending on the trait of choice.