Effects of genotype x environment interactions on the selection of broadly adapted Pacific oysters (Crassostrea gigas)

Broadly adapted genotypes are required if a single line of selection is to improve Pacific oyster (Crassostrea gigas) production throughout the heterogeneous growing environments found in the Pacific Northwest region of the United States. An experiment was conducted to determine if the relative rankings of average family performance (i.e. average individual weight, survival and yield) remain stable across a wide range of growing environments. Twenty-four full-sib oyster families were each planted both intertidally and subtidally in Yaquina Bay, Oregon, and Dabob Bay, Washington. Each family was represented by up to 12 replicate growout bags (stocked with 60 spat each) or 12 replicate lantern nets compartments (stocked with 40 spat each). Oysters were planted in Fall 2002 and yield (kg live weight replicate(-1)), average individual weight (g) and survival (%) were measured in Spring and Fall 2003 and at harvest in Summer 2004. Harvest average individual weight, survival, and yield were all significantly affected by family, environment, and family x environment interactions (P < 0.01). Family, environment, and family x environment interaction, accounted for 5%, 80%, and 2% of the total phenotypic variation in harvest average individual weight, respectively, 36%, 16%, and 7% of the total phenotypic variation in harvest survival, respectively, and 14%, 62%, and 5% of the total phenotypic variation in harvest yield, respectively. Although the effects of family x environment interactions were statistically significant, correlations among average family yields between all growout environments were significantly greater than 0 (r(p)> 0.397; P < 0.04) indicating high-yielding families in one environment tended to be high-yielding in other environments. Results of this study suggest that the effect of GxE interactions on average individual weight, survival, and yield, while significant, are not large enough to prevent selection in a limited number of well chosen environments from resulting in favorable correlated gains in other environments in the Pacific Northwest. (c) 2006 Elsevier B.V. All rights reserved.