Genetic parameters and genotype by environment interaction for body weight of Oreochromis shiranus

Genetic and phenotypic parameter estimates for tagging and harvest body weights were obtained from the F1 base (377 offspring of 59 full-sib families), F2 (1241 offspring of 52 full- and 33 half-sib families) and F3 (1052 offspring of 55 full- and 25 half-sib families) generations of Oreochromis shiranus. The fish were reared in earthen ponds at three different test environments in Malawi, which were high altitude (Chisitu), medium altitude (Domasi) and low altitude (Kasinthula) fish farms. The heritability estimate for tagging body weight was 0.31 ± 0.24, but likely inflated by the common full-sib effect. Heritability estimates for harvest body weight at each test environment were of medium magnitude (0.25 ± 0.07, 0.24 ± 0.11, 0.21 ± 0.06) for Chisitu, Domasi and Kasinthula, respectively. However, across generation and test environment, the heritability estimate was low (0.12 ± 0.07). The magnitude of the common full-sib effect for harvest body weight was substantial both within and across test environments. The genetic (r = 0.38 ± 0.13) and phenotypic (r = 0.46) correlations between tagging and harvest body weights were low. All genetic correlations between harvest body weights at the different test environments were not significantly different from unity, suggesting a low genotype by test environment interaction for the trait; high between Chisitu and Domasi (r = 0.95 ± 0.15), but moderate between Chisitu and Kasinthula (r = 0.74 ± 0.17) and between Domasi and Kasinthula (0.63 ± 0.20). However, more reliable genetic correlations between harvest body weights at the different test environments should be obtained using replicated ponds at each test environment before conclusions are drawn on the need to develop specific breeding programs for the diverse agro-climatic conditions of the Malawian small-scale pond aquaculture. Heritability estimate for the harvest body weight of males (0.40 ± 0.13) was higher than that of females (0.27 ± 0.08). The genetic correlation between the harvest body weights of males and females was high (r = 0.85 ± 0.15). The results show that there is quite substantial additive genetic variance for harvest body weight in O. shiranus that can be exploited through selective breeding.