Predicted response to selection for crossbred performance in mice.

Abstract

Compared in this study were predicted responses of three crossbred selection methods, using estimates of genetic parameters for seven quantitative traits from two genetically diverse mouse populations, designated C2 and ICR, and their reciprocal interpopulation crosses. The methods of crossbred selection were mass selection (MS), half-sib family selection (HS) and reciprocal recurrent selection (RRS). Traits studied were number born, 12-day litter weight, individual weights at birth, 21 (weaning), 42 and 56 days of age and gain from 21 to 42 days. Expected efficiency of MS was larger than HS for all traits. Expected efficiency of RRS was greater than MS for number born only. Average genetic and average maternal differences between the two populations were significant for all traits. Heterosis, measured as the mean difference between reciprocal F1 crosses and midparent values was also significant for all traits, varying from 3 to 6% of the respective trait means. Heritabilities in the purebred populations estimated from half-sib covariances were small for number born and weaning weight and moderately high for 12-day litter weight, post-weaning gain and adult body weights. Ratios of the sum of both purebred half-sib covariances to both crossbred half-sib covariances were greater than one for 12-day litter weight and the postweaning weight traits, which is compatible with a partial dominance gene model. This ratio was less than one for litter size, suggesting a possible overdominance situation. Genetic covariances and correlations between purebred and crossbred half-sib family means for the same trait generally agreed with the ratio estimates regarding inference to the type of gene action present.