Abstract

The rates of genetic gain and inbreeding were examined in alternative breeding designs of rainbow trout with different mating ratios, variable number of individuals measured and different number of traits included in a selection index in a closed nucleus-breeding scheme. Three body weight records during growth at the nucleus central station and body weight before marketing at a separate sea station (the breeding objective) were assumed to be recorded, and the genetic parameters were obtained from the actual Finnish breeding program. The rates of genetic gain were determined using the prediction error variance-covariance matrix of the traits included into the best linear prediction of breeding values. The rates of inbreeding were calculated using a first-order approximation, by empirically obtaining the probabilities of co-selection of relatives. The analysis showed that the rate of genetic gain can be improved as much as 20% by changing the mating ratios from the traditional nested designs (e.g., ratio of sires to dams 1M : 3F) to factorial mating (e.g., 3M : 5F). This enhance in genetic gain is mainly due to an increase in the selection intensity of females, which is constant in the nested designs with a fixed number of full-sib family tanks for a given family size. The rates of inbreeding appear to be higher for factorial than for nested designs, although, at the same rate of inbreeding, factorial designs present equal or higher rates of genetic gain compared to nested designs. The accuracy of the breeding values differ only little among the different mating ratios explored, whereas the inclusion of information from relatives at the sea station in the selection index increased the accuracy, and thus, the genetic gain. © 2005 Elsevier B.V. All rights reserved.