Two-way selection for muscle lipid content in pan-size rainbow trout (Oncorhynchus mykiss)

A two-way selection for muscle lipid content was performed in rainbow trout. One-year-old pan-size individuals were mass selected using Distell Fish Fatmeter values corrected for absolute weight (fat index, FI), this being a non-destructive measurement of the muscle lipid content in live fish. Direct and correlated responses at pan-size (around 260 g) were estimated after two generations of selection. Absolute Fat values and FI values in the upward line (fat line, FL) and downward line (lean line, LL) differed from the control line and from each other. The mean realized heritability achieved for FI was 0.25. The difference in Fat values between FL and LL resulted in a significant difference in muscle lipid content (29.6% dry matter content in FL, vs. 25.6% in LL, P=0.003). The differences in Fat and FI values between FL and LL were sustained until maturation at 2 years old. At the age of 1 year, the two lines did not differ significantly in terms of weight, length or most body shape traits (height, width or cross-section shape). However, FL fish had a better condition factor than LL (1.44 vs. 1.38), and a greater belly thickness (11% relative increase), the control line displaying intermediate values. The abdominal internal fat coat was thicker in FL, and represented a larger proportion of the entire abdominal thickness than in LL. The relative visceral weight (in % of total body weight), carcass and fillet yield were not modified by selection. The frequency of precocious males was significantly higher in FL (16.5%, vs. 11% and 10.1% in control and LL fish, respectively), but there was no difference between the reproductive traits of 2-year-old LL and FL females. These results indicate that breeding based on Fatmeter values was efficient in selecting the fillet lipid content in rainbow trout, with limited correlated responses at the size where selection was applied. In particular, no adverse effects were recorded of a reduction in muscle lipid content on carcass or fillet yield under the experimental conditions applied during this study.