Tidsskrift: Aquaculture, vol. 273, p. 33–41, 2007
Open Access: none
Realised and predicted responses to selection were obtained for harvest body weight of Oreochromis shiranus from two generations of selection at the National Aquaculture Centre, Domasi, Malawi. The realised response was estimated from the difference in the least squares means of selected and unselected control populations, while the predicted response was obtained from the difference of the mean breeding values between generations. The realised (13.2%) and predicted (14.4%) responses for harvest body weight over the two generations of selection were not significantly different (P>0.05). The intensity of selection from F-1 to F-2 (0.66) and from F-2 to F-3 (0.59) was low due to availability of few selection candidates per family at harvest caused by mortality, tag loss and the need to form broad heterogeneous multi-strain F-2 and F-3 populations. Within generation, the heritability estimates for the F-1 (0.31 +/- 0.09), F-2 (0.31 +/- 0.10) and F-3 (0.35 +/- 0.11) generations were moderate. Across generation, the heritability estimate was low (0.14 +/- 0.27). Within generation, the magnitude of the common full-sib effect was moderate (0.08-0.09) and not significantly different from zero (P> 0.05). However, across generations, the common full-sib effect was high (0.27 +/- 0.05) and significantly different from zero (P<0.05). A substantial correlated selection response due to selection for increased harvest body weight was observed for both tagging body weight and for survival from tagging to harvest; for tagging body weight 22.8% from F-1 to F-2 and 15.9% from F-2 to F-3; for survival 7.2% from F-1 to F-2 and 13.8% from F-2 to F-3. The results are discussed in relation to the maximisation of the genetic gain in the future generations while constraining the rate of inbreeding. (C) 2007 Elsevier B.V. All rights reserved.