Genetically, fish are characterised by a large variation in most traits compared to mammals. While the coefficient of variation is 5 to 10% in mammals it goes up to 30% in fish. This means, that even though a low heritability is found in many traits, the overall resulting genetic variation is large and improves the potential for selection in the fish population.
Fish are still largely undomesticated, and because relatively little is known about many basic features of their genetic constitution, genetic improvement studies have wider implications.
The main purpose of selection in the case of domesticated fish species is to increase the productivity of the existing and newly developed breeds (strains). Such an increase may be achieved by raising the growth rate, survival and numerous other quantitative traits in a cultured fish species. Many examples could be given to show large improvements in production traits. Danish investigations showed those significant improvements with respect to sustainability and environmental impact could be achieved through systematic breeding of rainbow trout families. Using the best performing family, it might be possible to produce more fish using the same amount of feed and concomitantly reduce the output of nitrogen and phosphorous to the environment. Such selection may prove to be a possibility for improving the food conversion efficiency.
Recent study results demonstrate that the progeny of stress-tolerant parents display improved performance under the conditions of aquaculture, relative to the progeny of unselected parents, in terms of growth. It has been demonstrated that their progeny inherits the stress-related traits identified in the parents, and the heritability of such selected traits has been quantified as moderately high.